阅读说明：本技术 一种信息预测方法、装置、设备和存储介质 (Information prediction method, device, equipment and storage medium ) 是由 王鑫 于 2021-10-26 设计创作，主要内容包括：本发明实施例公开了一种信息预测方法、装置、设备和存储介质,该方法包括：获取当前预测日期之前的历史时间段内目标物品的历史需求量和历史价值属性值；根据历史需求量和历史价值属性值确定目标物品对应的第一特征信息；将第一特征信息输入至预设回归树模型中,根据预设回归树模型的输出,确定当前预测日期内目标物品的预测需求量；根据历史需求量和历史价值属性值,确定当前预测日期内目标物品的预估价值属性值,并将当前预测日期的下一日期作为当前预测日期,返回执行获取当前预测日期之前的历史时间段内目标物品的历史需求量和历史价值属性值的操作,直到当前预测日期为目标预测日期为止,从而提高需求量信息预测的准确性。(The embodiment of the invention discloses an information prediction method, an information prediction device, information prediction equipment and a storage medium, wherein the method comprises the following steps: acquiring historical demand and historical value attribute values of a target item in a historical time period before a current forecast date; determining first characteristic information corresponding to the target object according to the historical demand and the historical value attribute value; inputting the first characteristic information into a preset regression tree model, and determining the predicted demand of the target object within the current prediction date according to the output of the preset regression tree model; and determining the estimated value attribute value of the target item in the current prediction date according to the historical demand and the historical value attribute value, taking the next date of the current prediction date as the current prediction date, and returning to execute the operation of acquiring the historical demand and the historical value attribute value of the target item in the historical time period before the current prediction date until the current prediction date is the target prediction date, thereby improving the accuracy of demand information prediction.)

1. An information prediction method, comprising:

acquiring historical demand and historical value attribute values of a target article in a historical time period before a current prediction date, wherein an initial value of the current prediction date is the current date;

determining first characteristic information corresponding to the target object according to the historical demand and the historical value attribute value;

inputting the first characteristic information into a preset regression tree model for demand information prediction, and determining the predicted demand of the target article within the current prediction date according to the output of the preset regression tree model;

and determining the estimated value attribute value of the target item in the current prediction date according to the historical demand and the historical value attribute value, taking the next date of the current prediction date as the current prediction date, and returning to execute the operation of acquiring the historical demand and the historical value attribute value of the target item in the historical time period before the current prediction date until the current prediction date is the target prediction date.

2. The method of claim 1, wherein determining the projected value attribute value for the target item within the current forecast date based on the historical demand and the historical value attribute value comprises:

determining a demand explosion coefficient corresponding to a current prediction date according to historical demands corresponding to various articles belonging to a target category, wherein the target category is the category to which the target article belongs;

acquiring a month granularity regression coefficient corresponding to the target object;

based on a predetermined month granularity discount proportion corresponding to the target article, determining a target discount proportion corresponding to a month previous to a target month in which the current prediction date is located;

determining a value attribute value discount ratio between the target month and the previous month according to the demand burst coefficient, the month granularity regression coefficient and the target discount ratio;

and determining the estimated value attribute value of the target item in the current prediction date according to the value attribute value discount proportion and the first value attribute value of the target item in the previous month.

3. The method according to claim 2, wherein the obtaining the month-granularity regression coefficient corresponding to the target item comprises:

determining a month granularity discount proportion corresponding to the target object according to the historical demand, the historical value attribute value and the historical object acquisition cost corresponding to each object belonging to the target category;

determining the total monthly granularity article acquisition cost corresponding to the target article according to the historical demand and the historical article acquisition cost corresponding to each article belonging to the target category;

and performing linear regression on the month granularity discount proportion and the total acquired cost of the month granularity articles, and determining a month granularity regression coefficient corresponding to the target article.

4. The method of claim 2, wherein said determining a projected value attribute value for said target item on a current forecast date based on said value attribute value discount rate and a first value attribute value for said target item in said previous month of January comprises:

determining a value attribute value pre-estimation coefficient corresponding to the target month according to the value attribute value discount proportion, the first value attribute value of the target item in the previous month and the second value attribute value of the target item in the target month in a historical time period;

and determining the estimated value attribute value of the target object in the current prediction date according to the value attribute value estimation coefficient and a third value attribute value corresponding to the history date which is in the same period as the current prediction date in the history time period.

5. The method of claim 1, wherein the first feature information comprises: demand statistics, value attribute value statistics, and cross-over features between demand and value attribute values.

6. The method according to claim 5, wherein the determining first characteristic information corresponding to the target item according to the historical demand amount and the historical value attribute value comprises:

determining a first variance according to the historical demand and determining a second variance according to the historical value attribute value;

determining covariance between the demand and the value attribute value according to the historical demand and the historical value attribute value;

and determining the cross feature between the demand and the value attribute value according to the first variance, the second variance and the covariance.

7. The method of claim 1, prior to using the pre-set regression tree model, further comprising:

and training the preset regression tree model according to the sample characteristic information and the actual demand quantity based on a gradient descent mode.

8. The method according to any one of claims 1-7, further comprising:

acquiring historical browsing quantity of a target article in a historical time period before a current prediction date, and determining second characteristic information corresponding to the target article according to the historical browsing quantity;

the step of inputting the first characteristic information into a preset regression tree model for demand information prediction, and determining the predicted demand of the target item within the current prediction date according to the output of the preset regression tree model includes:

inputting the first characteristic information and the second characteristic information into a preset regression tree model for demand information prediction, and determining the predicted demand of the target object within the current prediction date according to the output of the preset regression tree model;

after the determining the predicted value attribute value of the target item within the current forecast date, the method further comprises the following steps:

and determining the estimated browsing amount of the target object within the current prediction date according to the historical demand and the historical browsing amount.

9. The method of claim 8, wherein the second feature information comprises: the statistical characteristics of the browsing amount and the cross characteristics between the demand amount and the browsing amount;

the determining second characteristic information corresponding to the target item according to the historical browsing amount includes:

determining the statistical characteristics of the browsing amount according to the historical browsing amount;

and determining the cross characteristics between the demand and the browsing amount according to the historical browsing amount and the historical demand.

10. The method of claim 8, wherein determining the estimated view volume of the target item within the current forecast date based on the historical demand volume and the historical view volume comprises:

determining a first historical demand total amount of a target year of a current forecast date and a second historical demand total amount of the target year in the same time period of a previous year according to the historical demand amount;

determining a demand increase proportion according to the first historical demand total amount and the second historical demand total amount;

and determining the estimated browsing amount of the target item in the current prediction date according to the demand increase proportion and the historical browsing amount corresponding to the historical date in the historical time period which is in the same period as the current prediction date.

11. The method of claim 1, after determining the predicted demand for the target item within the target prediction date, further comprising:

acquiring a target prediction demand corresponding to each day in preset spare days after the current day;

determining a target inventory corresponding to the target object according to the target forecast demand, a corresponding numerical value of a preset spot rate level under normal distribution and a preset standard deviation;

and determining the target replenishment quantity according to the target inventory quantity and the current inventory quantity.

12. An information prediction apparatus, comprising:

the historical data acquisition module is used for acquiring the historical demand and the historical value attribute value of the target object in a historical time period before the current prediction date, wherein the initial value of the current prediction date is the current date;

the first characteristic information determining module is used for determining first characteristic information corresponding to the target object according to the historical demand and the historical value attribute value;

the information prediction module is used for inputting the first characteristic information into a preset regression tree model to predict demand information and determining the predicted demand of the target object within the current prediction date according to the output of the preset regression tree model;

and the estimated value attribute value determining module is used for determining the estimated value attribute value of the target item in the current prediction date according to the historical demand and the historical value attribute value, taking the next date of the current prediction date as the current prediction date, and returning to execute the operation of acquiring the historical demand and the historical value attribute value of the target item in the historical time period before the current prediction date until the current prediction date is the target prediction date.

13. An electronic device, characterized in that the electronic device comprises:

one or more processors;

a memory for storing one or more programs;

when executed by the one or more processors, cause the one or more processors to implement the information prediction method of any one of claims 1-11.

14. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the information prediction method according to any one of claims 1 to 11.

Technical Field

Embodiments of the present invention relate to computer technologies, and in particular, to an information prediction method, apparatus, device, and storage medium.

Background

With the rapid development of the internet, more and more people like online shopping. Generally, there are many fixed sales promotions on a shopping platform every year to increase the sales of items, i.e. the demand for items.

Currently, demand information during a subsequent promotion is often predicted based on historical demand during a previous promotion, so that advance stock can be made based on the prediction.

However, in the process of implementing the present invention, the inventor finds that at least the following problems exist in the prior art:

the existing method only carries out information prediction based on historical demand, neglects the information such as price causing the root of the demand change, namely the information such as value attribute value, and the like, thereby causing the predicted demand information to be inaccurate, and further causing the condition of insufficient stock or overstock goods.

Disclosure of Invention

The embodiment of the invention provides an information prediction method, an information prediction device, information prediction equipment and a storage medium, and aims to improve the accuracy of demand information prediction.

In a first aspect, an embodiment of the present invention provides an information prediction method, including:

acquiring historical demand and historical value attribute values of a target article in a historical time period before a current prediction date, wherein an initial value of the current prediction date is the current date;

determining first characteristic information corresponding to the target object according to the historical demand and the historical value attribute value;

inputting the first characteristic information into a preset regression tree model for demand information prediction, and determining the predicted demand of the target article within the current prediction date according to the output of the preset regression tree model;

and determining the estimated value attribute value of the target item in the current prediction date according to the historical demand and the historical value attribute value, taking the next date of the current prediction date as the current prediction date, and returning to execute the operation of acquiring the historical demand and the historical value attribute value of the target item in the historical time period before the current prediction date until the current prediction date is the target prediction date.

In a second aspect, an embodiment of the present invention further provides an information prediction apparatus, including:

the historical data acquisition module is used for acquiring the historical demand and the historical value attribute value of the target object in a historical time period before the current prediction date, wherein the initial value of the current prediction date is the current date;

the first characteristic information determining module is used for determining first characteristic information corresponding to the target object according to the historical demand and the historical value attribute value;

the information prediction module is used for inputting the first characteristic information into a preset regression tree model to predict demand information and determining the predicted demand of the target object within the current prediction date according to the output of the preset regression tree model;

and the estimated value attribute value determining module is used for determining the estimated value attribute value of the target item in the current prediction date according to the historical demand and the historical value attribute value, taking the next date of the current prediction date as the current prediction date, and returning to execute the operation of acquiring the historical demand and the historical value attribute value of the target item in the historical time period before the current prediction date until the current prediction date is the target prediction date.

In a third aspect, an embodiment of the present invention further provides an electronic device, where the electronic device includes:

one or more processors;

a memory for storing one or more programs;

when executed by the one or more processors, cause the one or more processors to implement an information prediction method as provided by any of the embodiments of the invention.

In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to implement the information prediction method according to any embodiment of the present invention.

The embodiment of the invention has the following advantages or beneficial effects:

the current date is used as an initial value of the current prediction date, the historical demand and the historical value attribute value of the target object in the historical time period before the current prediction date are obtained, the first characteristic information corresponding to the target object is determined according to the historical demand and the historical value attribute value, the first characteristic information is input into the preset regression tree model for demand information prediction, the predicted demand of the target object in the current prediction date is determined according to the output of the preset regression tree model, and therefore the preset regression tree model can predict demand information based on the historical demand and the historical value attribute value at the same time, and accuracy of demand information prediction is improved. And when the current prediction date does not reach the target prediction date, determining the predicted value attribute value of the target object in the current prediction date according to the historical demand and the historical value attribute value, and returning and executing the operation by taking the next date of the current prediction date as the current prediction date, so that the predicted value attribute value can be used as the historical value attribute value of the current prediction date to determine the predicted demand of the target object in the next prediction date, thereby accurately obtaining the predicted demand of the target object in the target prediction date in a circular iteration mode, and improving the accuracy of demand information prediction.

Drawings

Fig. 1 is a flowchart of an information prediction method according to an embodiment of the present invention;

FIG. 2 is an example of a single tree training process in a predictive regression tree model according to an embodiment of the present invention;

FIG. 3 is a flowchart of an information prediction method according to a second embodiment of the present invention;

fig. 4 is a flowchart of an information prediction method according to a third embodiment of the present invention;

fig. 5 is a schematic structural diagram of an information prediction apparatus according to a fourth embodiment of the present invention;

fig. 6 is a schematic structural diagram of an electronic device according to a fifth embodiment of the present invention.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.

Example one

Fig. 1 is a flowchart of an information prediction method according to an embodiment of the present invention, which is applicable to a case of predicting demand information of an article in any prediction date. The method may be performed by an information prediction apparatus, which may be implemented by software and/or hardware, integrated in an electronic device. As shown in fig. 1, the method specifically includes the following steps:

s110, acquiring the historical demand and the historical value attribute value of the target item in the historical time period before the current prediction date, wherein the initial value of the current prediction date is the current date.

The target item may be an item of SKU (Stock Keeping Unit) granularity of the demand to be predicted. The current forecast date may refer to a date on which the demand amount needs to be forecasted currently. The current forecast date is dynamically updated, that is, the current forecast date is updated once per cycle, so as to forecast the demand of the target item in different dates. The history time period may refer to a history date with the current prediction date as a time node. The historical time period may refer to one week, one month, two months, etc. of history. For example, if the current date is 21 years, 10 months, and 13 days, the historical week may be 21 years, 10 months, and 6 days to 21 years, 10 months, and 12 days, and the historical contemporaneous month may be 20 years, 10 months, and 6 days to 20 years, 10 months, and 12 days. The historical demand for the target item may refer to the daily sales of the target item over a historical period of time. The historical value attribute value for the target item may refer to the daily price of the target item over a historical period of time.

Specifically, the current date may be taken as an initial value of the current prediction date to perform a first loop, so as to predict the predicted demand amount of the target item within the current date for the first time.

And S120, determining first characteristic information corresponding to the target object according to the historical demand and the historical value attribute value.

Specifically, the obtained historical demand and the historical value attribute value may be subjected to statistical processing, so as to obtain a demand statistical characteristic and a value attribute value statistical characteristic. For example, the historical demand of each day in the historical week, the historical month, the historical two months, the historical contemporaneous month and the historical two months can be respectively counted to determine the average value of the historical demands, the maximum value of the historical demands, the minimum value of the historical demands, the variance of the historical demands, the median of the historical demands and the regression coefficient. The regression coefficient may be a regression coefficient obtained by performing linear regression on each historical demand. The ratio of the daily historical demand in the historical week to the median of the historical demand corresponding to the historical 1 month can also be used as a demand statistical characteristic to measure the change condition of the historical demand.

It should be noted that the backorder filling and smoothing operations may be performed on the historical demand when determining the statistical characteristics of the demand in order to ensure the accuracy of the statistical characteristics.

Illustratively, the first feature information may include, in addition to the demand statistic feature and the value attribute value statistic feature, a cross feature between the demand and the value attribute value to measure a variation relationship between the demand and the value attribute value. For example, S120 may include: determining a first variance according to the historical demand and determining a second variance according to the historical value attribute value; determining covariance between the demand and the value attribute value according to the historical demand and the historical value attribute value; and determining the cross feature between the demand and the value attribute value according to the first variance, the second variance and the covariance. Specifically, for each of the historical time periods, a first variance var (X) corresponding to the demand may be determined based on each historical demand, a second variance var (Y) corresponding to the value attribute may be determined based on each historical value, and a covariance cov (X, Y) between the demand and the value attribute may be determined based on each historical demand and each historical value, and may be determined by a formulaThe cross signature corr between the demand and the value of the value attribute is determined. The accuracy of information prediction is further improved by determining the cross characteristics between the demand and the value attribute values.

For example, the first characteristic information may further include target item basic information to further improve the accuracy of information prediction. For example, the target item basis information may be, but is not limited to, the name and category of the target item. The category may refer to the brand and/or class of the target item.

S130, inputting the first characteristic information into a preset regression tree model for demand information prediction, and determining the predicted demand of the target object within the current prediction date according to the output of the preset regression tree model.

The preset regression tree model may refer to a decision tree model for regression prediction. The preset regression tree model in this embodiment may be obtained by training in advance based on the sample feature information and the actual demand.

Specifically, the embodiment may combine the target item basic information b in the first feature information₁Statistical characteristics of demand s₁Value attribute value statistical characteristic p₁And cross-feature sp between demand and value attribute values₁Characteristic vector inputData capable of being received by conversion model_predic＝X₁＝[b₁s₁p₁sp₁]And inputting the data into a trained preset regression tree model to predict the demand information. The preset regression tree model can output the predicted demand, namely P ═ f_{inputDatatrain}(inputData_predic) Thereby obtaining the predicted demand amount P of the target item within the current prediction date. In the embodiment, the value attribute value information causing the root of the change of the demand is input into the preset regression tree model, so that the preset regression tree model can predict the demand information based on the historical demand and the historical value attribute value at the same time, and the accuracy of predicting the demand information can be improved.

For example, the first feature information corresponding to each article to be predicted may be subjected to a stitching process to obtain a feature matrix, such asAnd the characteristic matrix is input into the preset regression tree model, so that the preset regression tree model can predict the demand information of a plurality of articles at the same time, and the prediction efficiency is improved.

S140, detecting whether the current predicted date is the target predicted date, if not, executing step S150, and if so, executing step S160.

The target prediction date may refer to a date to be predicted finally or a last date in a prediction time period. For example, if the demand amount for 11/2021 is predicted, 11/2021 may be set as the target predicted date. Alternatively, if the demand amount from 11/1/2021 to 11/2021 is predicted, 11/2021/11 may be set as the target predicted date.

Specifically, since the current prediction date is dynamically updated along with the loop operation, it is necessary to detect whether the current prediction date is the target prediction date to determine whether the prediction operation is completed.

And S150, determining the estimated value attribute value of the target item in the current prediction date according to the historical demand and the historical value attribute value, taking the next date of the current prediction date as the current prediction date, and returning to execute the operation of S110.

The predicted value attribute value may refer to a value attribute value of the estimated target item within the current prediction date. The predicted value attribute value may refer to a predicted price of the target item. It should be noted that, since the price of the article on the future date belongs to the confidential information and cannot be known in advance, the price of the article on the future date needs to be estimated.

Specifically, when the current forecast date is not the target forecast date, indicating that the target forecast date has not been reached, and the forecasting operation has not ended, the estimated value attribute value of the target item in the current forecast date may be estimated based on the historical demand and the historical value attribute value, so that demand information forecasting of the subsequent forecast date may be performed based on the forecast demand and the estimated value attribute value corresponding to the current forecast date. By setting the next date of the current predicted date as the current predicted date and returning to perform the operation of S110, the predicted required amount corresponding to the next predicted date can be determined by the next loop. When the predicted demand corresponding to the next prediction date is predicted, the predicted demand and the estimated value attribute value corresponding to the current prediction date can be used as the historical demand and the historical value attribute value corresponding to the current prediction date to perform information prediction, so that the demand information prediction can be performed on the basis of the historical demand and the historical value attribute value corresponding to the recent historical time period in each cycle, and the accuracy of demand information prediction is guaranteed.

For example, if the current date is 21 years, 10 months and 13 days, and the demand of 2021 years, 11 months and 11 days is to be predicted, the first loop may be performed with 21 years, 10 months and 13 days as the current prediction date, the predicted demand corresponding to 21 years, 10 months and 13 days is determined, then the estimated value attribute value corresponding to 21 years, 10 months and 13 days is estimated, and the second loop may be performed with 21 years, 10 months and 14 days as the current prediction date, so that information prediction is performed based on the historical demand and the historical value attribute value (including the predicted demand and the estimated value attribute value corresponding to 21 years, 10 months and 13 days) before 21 years, 10 months and 14 days, the predicted demand corresponding to 21 years, 10 months and 14 days is obtained, and prediction is performed day by day sequentially until the predicted demand corresponding to 2021 years, 11 months and 11 days is predicted. The predicted demand of the target article in the target prediction date can be more accurately obtained in a loop iteration mode, and the accuracy of demand information prediction is improved.

And S160, stopping the prediction operation.

According to the technical scheme of the embodiment, the current date is used as the initial value of the current prediction date, the historical demand and the historical value attribute value of the target object in the historical time period before the current prediction date are obtained, the first characteristic information corresponding to the target object is determined according to the historical demand and the historical value attribute value, the first characteristic information is input into the preset regression tree model to predict the demand information, the predicted demand of the target object in the current prediction date is determined according to the output of the preset regression tree model, and therefore the preset regression tree model can predict the demand information based on the historical demand and the historical value attribute value at the same time, and the accuracy of demand information prediction is improved. And when the current prediction date does not reach the target prediction date, determining the predicted value attribute value of the target object in the current prediction date according to the historical demand and the historical value attribute value, and returning and executing the operation by taking the next date of the current prediction date as the current prediction date, so that the predicted value attribute value can be used as the historical value attribute value of the current prediction date to determine the predicted demand of the target object in the next prediction date, thereby accurately obtaining the predicted demand of the target object in the target prediction date in a circular iteration mode, and improving the accuracy of demand information prediction.

On the basis of the above technical solution, before using the preset regression tree model, the method may further include: and training the preset regression tree model according to the sample characteristic information and the actual demand quantity based on a gradient descent mode.

The sample feature information may refer to feature information corresponding to a historical prediction date. The actual demand may refer to the actual demand corresponding to the historical forecast date. The sample characteristic information may be determined based on the first characteristic information acquisition manner described above.

Specifically, fig. 2 shows an example of a single tree training process in a pre-set regression tree model. As shown in FIG. 2, sample feature information { X ] may be input_iRandomly selecting M characteristics to search for the optimal cutting point, and the cutting modes (i.e. whether the cutting modes are the same) of the discrete characteristics 0 and 1 are shown in FIG. 2. Continuity variables such as [1,100 ]]The optimal cut point in this interval needs to be found during training to minimize the loss function, i.e. in [1,100 ]]Find a value that minimizes the loss function, i.e., the best cut point. The loss function in this embodiment may adopt a mixed mean Square error mse (mean Square error) function. For example, the penalty function for the t-th tree is as follows:

wherein the content of the first and second substances,is the real demand;model parameters for the t-1 th tree,input feature information for the t-1 st tree,is the predicted result of the t-1 th tree. Each tree in the gradient descent tree model learns the residuals of the result sums of all previous trees. By Taylor second order expansion of the above, i.e.The above loss function can be converted into:wherein, g_iIs a first derivative coefficient, h_iIs a second derivative coefficient. With the training of the pre-set regression tree model, i.e., the increase of the tree, the predicted demand will gradually approach the actual demand. Meanwhile, the fitting condition of the model can be supervised by using the test data set, the model is continuously close to the true value in the continuous iteration process, the prediction effect on the test data set is not deteriorated, the training can be continuously carried out until the prediction effect of the test data set is not improved, and the training of the preset regression tree model is finished.

On the basis of the above technical solutions, after S160, the method further includes: acquiring a target prediction demand corresponding to each day in preset spare days after the current day; determining a target inventory corresponding to the target object according to the target forecast demand, the corresponding numerical value of the preset spot rate level under normal distribution and the preset standard deviation; and determining the target replenishment quantity according to the target inventory quantity and the current inventory quantity.

The preset stock-keeping days can be preset days needing stock keeping. The preset spot rate level may be a preset service level value for the item. The spot rate level follows a normal distribution. For example, if the current cargo rate level is 90%, the corresponding value of the 90% current cargo rate level under the normal distribution can be obtained by looking up the table. The preset standard deviation may be a standard deviation of a difference between a historical predicted demand amount and an actual demand amount determined in advance. The historical predicted demand may be a predicted demand for historical dates. The actual demand may refer to the actual demand on the history date. The current inventory amount may refer to the existing inventory amount of the target item at the distribution center.

Specifically, the predicted demand amount corresponding to each predicted date may be stored in a hive (data warehouse tool) data table, so that the replenishment device may obtain the predicted demand amount from the hive data table through a hive interface. The target predicted demand amount corresponding to each day in the preset equipment days after the current date may be acquired based on the predicted demand amount corresponding to each predicted date. For example, if the current date is 21 years, 10 months, and 13 days, and the preset stock-keeping days are 20 days, the target predicted demand amount per day in the last 20 days from 21 years, 10 months, and 13 days can be obtained. The obtained target forecast demand quantities can be added to obtain a target forecast demand total quantity, a numerical value corresponding to a preset current rate level under normal distribution is multiplied by a preset standard deviation to obtain an error quantity, the target forecast demand total quantity and the error quantity are added to obtain a target stock quantity corresponding to a target article, and a difference value between the target stock quantity and the current stock quantity is used as a target replenishment quantity, so that the more accurate replenishment quantity of a distribution center where the target article is located can be obtained. In this embodiment, the obtained target replenishment quantity may be stored for the purchase order creating device in the downstream to call, so that the purchase order creating device may create a purchase order based on the obtained target replenishment quantity to perform the purchase stock.

It should be noted that the magnitude of the predicted demand obtained in this embodiment can meet the actual demand in the sales promotion stage, so that the stock parameters in the replenishment device do not need to be modified manually in addition in the stock stage, the work of the user for adjusting the stock days for stock is omitted, and further more accurate stock quantity can be obtained, and finally, the stock quantity can be delivered to the consumer more efficiently.

Example two

Fig. 3 is a flowchart of an information prediction method according to a second embodiment of the present invention, and in this embodiment, based on the above embodiments, further optimization is performed on the step of determining the estimated value attribute value of the target item in the current prediction date according to the historical demand and the historical value attribute value, where explanations of terms that are the same as or corresponding to the above embodiments are not repeated here.

Referring to fig. 3, the information prediction method provided in this embodiment specifically includes the following steps:

s310, acquiring the historical demand and the historical value attribute value of the target item in the historical time period before the current prediction date, wherein the initial value of the current prediction date is the current date.

S320, determining first characteristic information corresponding to the target object according to the historical demand and the historical value attribute value.

S330, inputting the first characteristic information into a preset regression tree model for demand information prediction, and determining the predicted demand of the target article in the current prediction date according to the output of the preset regression tree model.

S340, detecting whether the current predicted date is the target predicted date, if not, performing step S350, and if so, performing step S392.

And S350, determining a demand explosion coefficient corresponding to the current prediction date according to the historical demand corresponding to each article belonging to the target category, wherein the target category is the category to which the target article belongs.

Wherein the target category may refer to a brand and/or category to which the target item belongs.

Specifically, the historical demand of each item belonging to the target category in the historical time period may be acquired, for example, the historical demand of the last three years may be acquired, and the historical demand may be divided into the pre-promotion monthly demand and the promotion meso-monthly demand according to the current forecast date. The sales promotion previous month demand may refer to a demand of a month previous to a target month on which the current forecast date is located. The promotional monthly demand amount may refer to the demand amount for the month the current forecast date is in. For example, if the current forecast date is 21 years, 11 months and 11 days, the historical demand amounts of each item in 21 years and 10 months may be added, and the total demand amount of the month corresponding to 21 years and 10 months may be obtained as the demand amount of the month before sales promotion, and similarly, the total demand amount of the month corresponding to 20 years and 10 months may also be obtained as the demand amount of the month before sales promotion. And averaging the monthly demand of each promotion to obtain the monthly demand of the promotion. Adding the historical demand quantities of each article in 11 months in 20 years to obtain the monthly demand total quantity corresponding to 11 months in 20 years as the sales promotion monthly demand quantity, obtaining the monthly demand total quantity corresponding to 11 months in 19 years as the sales promotion monthly demand quantity in the same way, and averaging the sales promotion monthly demand quantities to obtain the sales promotion monthly demand quantity average value. And determining the ratio of the average value of the monthly demand in the promotion to the average value of the monthly demand before the promotion as the demand explosion coefficient corresponding to the current prediction date.

S360, obtaining a month granularity regression coefficient corresponding to the target object;

the month granularity regression coefficient may be a regression coefficient obtained by performing linear regression on the month granularity discount proportion and the total month granularity article acquisition cost, which are predetermined in months.

Exemplarily, S360 may include: determining a month granularity discount proportion corresponding to the target object according to the historical demand, the historical value attribute value and the historical object acquisition cost corresponding to each object belonging to the target category; determining the total monthly granularity article acquisition cost corresponding to the target article according to the historical demand and the historical article acquisition cost corresponding to each article belonging to the target category; and performing linear regression on the month granularity discount proportion and the total acquired cost of the month granularity articles, and determining a month granularity regression coefficient corresponding to the target article.

Wherein, the historical value attribute value may refer to a transaction price of the item. The historical item acquisition cost may refer to an item's shipping price. The total cost for acquiring a monthly granularity item may refer to monthly granularity GMV (Gross merchandisc Volume).

Specifically, for each month, the historical demand, the historical value attribute value, and the historical item acquisition cost of each item belonging to the target category in the current month may be obtained. And multiplying the historical demand of each article in the current month by the historical value attribute value to obtain the granularity cost of the first month corresponding to each article, and adding the granularity costs of the first month corresponding to each article to obtain the total granularity cost of the first month. In a similar way, the historical demand of each article in the current month is multiplied by the historical article acquisition cost to obtain a second February granularity cost corresponding to each article, and the second February granularity costs corresponding to each article are added to obtain a second February granularity total cost. And taking the ratio of the total monthly granularity cost to the total february granularity cost as the monthly granularity discount ratio of the target article in the current month. Similarly, a month-granular discount rate for the target item per month may be determined. For each month, multiplying the historical demand of each article in the current month by the historical article acquisition cost to obtain the month-granularity article acquisition cost corresponding to each article, and adding the month-granularity article acquisition costs corresponding to each article to obtain the total month-granularity article acquisition cost of the target article in the current month. Similarly, the total cost of the target item for acquiring the items at the month granularity of each month can be determined. And performing linear regression based on the month granularity discount proportion corresponding to each month and the total acquired cost of the month granularity articles, namely the total acquired cost of the month granularity articles is a multiplied by the month granularity discount proportion + b, and determining month granularity regression coefficients a (independent variable coefficients) and b (intercept terms) corresponding to the target articles by using a least square method.

S370, based on the predetermined month granularity discount proportion corresponding to the target article, determining the target discount proportion corresponding to the month before the target month where the current forecast date is located.

Specifically, the month-granularity discount rate of the target object in each month can be predetermined through the above-mentioned manner. In this embodiment, the month particle size discount ratio of the previous month of the target month in which the target item in the same year is located on the current prediction date may be used as the target discount ratio, or the month particle size discount ratios of the previous months in different years may be averaged, and the obtained average value of the month particle size discount ratios is used as the target discount ratio. For example, if the current prediction date is 21 years, 11 months and 11 days, the month-size discount rate corresponding to 21 years and 10 months may be used as the target discount rate, or the month-size discount rate corresponding to 10 months of the historical year may be averaged, and the obtained average value may be used as the target discount rate.

S380, determining the value attribute value discount proportion between the target month and the previous month according to the demand explosion coefficient, the month granularity regression coefficient and the target discount proportion.

Specifically, the value attribute value discount ratio between the target month and the previous month of January may be determined by the following formula:

wherein, K₁Is the value attribute value discount ratio between the target month and the previous month; k₂Is the demand explosion factor; k₃Is the target discount rate; a and b are monthly granularity regression coefficients.

S390, determining the estimated value attribute value of the target item in the current prediction date according to the value attribute value discount proportion and the first value attribute value of the target item in the previous month.

The first value attribute value may be a value attribute value corresponding to a previous month in a recent year, or may be an average value of value attribute values corresponding to previous months in historical years. For example, if the current prediction date is 11/21/year and the previous month of the target month in which the current prediction date is located is 10/month, the month price (average value of daily prices) of 10/21/year may be directly used as the first value attribute value, or the month price of 10/19/year, the month price of 10/20/year, and the month price of 10/21/year may be averaged to obtain the average value as the first value attribute value.

Specifically, the value attribute value discount ratio first value attribute value may be multiplied, the obtained multiplication result is the estimated value attribute value of the target item in the target month, and at this time, the estimated value attribute value may be directly used as the estimated value attribute value of the target item in the current prediction date.

Exemplarily, S390 may include: determining a value attribute value estimation coefficient corresponding to the target month according to the value attribute value discount proportion, the first value attribute value of the target item in the previous month and the second value attribute value of the target item in the target month in the historical time period; and determining the estimated value attribute value of the target object in the current prediction date according to the value attribute value estimation coefficient and the third value attribute value corresponding to the history date which is in the same period as the current prediction date in the history time period.

The second value attribute value may be a value attribute value corresponding to a target month in a previous year, or may be an average value of value attribute values corresponding to target months in historical years. For example, if the current prediction date is 21 years, 11 months and 11 days, and the target month in which the current prediction date is located is 11 months, the month price (average value of the daily prices) of 20 years and 11 months may be used as the second-value attribute value, or the month price of 18 years and 11 months, the month price of 19 years and 11 months, and the month price of 20 years and 11 months may be averaged to obtain the average value as the second-value attribute value. The third value attribute value may be a value attribute value on the same date as the current prediction date in the last year, or may be a mean value of value attribute values on the same date as the current prediction date in the past years. For example, if the current predicted date is 21 years, 11 months and 11 days, and the target month in which the current predicted date is located is 11 months, the daily price for 11 days of 11 months and 20 years may be used as the third-value attribute value, or the daily price for 11 days of 11 months and 18 years, the daily price for 11 days of 11 months and 19 years, and the daily price for 11 days of 11 months and 20 years may be averaged to obtain the average value as the third-value attribute value.

Specifically, the product obtained by multiplying the discount ratio of the value attribute value by the attribute value of the first value is divided by the attribute value of the second value, and the obtained division result is used as the estimation coefficient of the value attribute value corresponding to the target month, that is, the estimation price coefficient of the target month. And taking the product of the value attribute value estimation coefficient corresponding to the target month and the third value attribute value as the estimated value attribute value of the target object in the current prediction date, so that the estimated value attribute value corresponding to the current prediction date can be estimated more accurately in a day granularity mode.

S391, the operation of S310 is executed by returning the date next to the current predicted date.

S392, stopping the prediction operation.

According to the technical scheme of the embodiment, the discount proportion of the value attribute value between the target month and the previous month can be determined according to the demand explosion coefficient corresponding to the current prediction date, the month granularity regression coefficient corresponding to the target article and the target discount proportion corresponding to the previous month of the target month in which the current prediction date is located, and the estimated value attribute value of the target article in the current prediction date can be determined according to the discount proportion of the value attribute value and the first value attribute value of the target article in the previous month, so that the value attribute value of the future date can be accurately estimated, and the accuracy of information prediction is further improved.

EXAMPLE III

Fig. 4 is a flowchart of an information prediction method according to a third embodiment of the present invention, where in this embodiment, on the basis of the above-described embodiment, steps "obtaining a historical browsing amount of a target item in a historical time period before a current prediction date, and determining second feature information corresponding to the target item according to the historical browsing amount" are added, and on the basis, information prediction is performed based on a historical demand amount, a historical value attribute value, and a historical browsing amount. Wherein explanations of the same or corresponding terms as those of the above embodiments are omitted.

Referring to fig. 4, the information prediction method provided in this embodiment specifically includes the following steps:

s410, acquiring the historical demand and the historical value attribute value of the target item in the historical time period before the current prediction date, wherein the initial value of the current prediction date is the current date.

And S420, determining first characteristic information corresponding to the target article according to the historical demand and the historical value attribute value.

S430, obtaining the historical browsing amount of the target object in the historical time period before the current prediction date, and determining second characteristic information corresponding to the target object according to the historical browsing amount.

The browsing amount may refer to the number of users browsing the item information. For example, the browsing volume may refer to traffic volume. The historical browsing volume of the target item may refer to the daily traffic of the target item over a historical period of time.

Illustratively, the second characteristic information may include a browsing volume statistical characteristic. For example, S430 may include: and determining the statistical characteristics of the browsing amount according to the historical browsing amount. Specifically, the obtained historical browsing amount may be statistically processed to determine the statistical characteristics of the browsing amount. For example, the historical browsing amount per day in the historical week, the historical month, the historical two months, the historical contemporaneous month and the historical two months can be counted respectively, and the average value of the historical browsing amount, the maximum value of the historical browsing amount, the minimum value of the historical browsing amount, the variance of the historical browsing amount, the median of the historical browsing amount and the regression coefficient can be determined. The ratio of the daily historical browsing amount in the historical week to the median of the historical browsing amount corresponding to 1 month in the history can also be used as a browsing amount statistical characteristic to measure the change condition of the historical browsing amount.

Illustratively, the second characteristic information may include, in addition to the statistical characteristics of the browsing amount, cross characteristics between the demand amount and the browsing amount so as to measure a variation relationship between the demand amount and the browsing amount. For example, S430 may further include: and determining the cross characteristics between the demand and the browsing amount according to the historical browsing amount and the historical demand. Specifically, a first variance is determined according to the historical demand, and a third variance is determined according to the historical browsing amount; browsing based on historical demand and historyQuantity, determining the covariance between the demand quantity and the browsing quantity; and determining the cross feature between the demand and the browsing amount according to the first variance, the third variance and the covariance between the demand and the browsing amount. For example, for each of the historical time periods, a first variance var (X) corresponding to the demand amount may be determined based on each historical demand amount, a third variance var (Z) corresponding to the browsing amount may be determined based on each historical browsing amount, and a covariance cov (X, Z) between the demand amount and the browsing amount may be determined based on each historical demand amount and each historical browsing amount, and may be expressed by a formulaAnd determining the cross feature corr between the demand quantity and the browsing quantity. By determining the cross characteristics between the demand and the browsing volume, the accuracy of information prediction can be further improved.

Here, the execution sequence of step S430 is not limited. For example, step S430 may be performed sequentially after step S420, may be performed before step S420, and may be performed simultaneously with step S420.

S440, inputting the first characteristic information and the second characteristic information into a preset regression tree model for demand information prediction, and determining the predicted demand of the target object in the current prediction date according to the output of the preset regression tree model.

Specifically, the embodiment may combine the target item basic information b in the first feature information₁Statistical characteristics of demand s₁Value attribute value statistical characteristic p₁Cross-feature sp between demand and value attribute values₁And a browsing amount statistical characteristic l in the second characteristic information₁And cross-feature sl between demand and browsing volume₁Characteristic vector inputData capable of being received by conversion model_predic＝X₁＝[b₁s₁p₁l₁sp₁sl₁]Inputting the value attribute value information and the browsing amount information into a preset regression tree model after training for demand information prediction, thereby inputting the value attribute value information and the browsing amount information which cause the change root of the demand into a preset regression tree modelIn the regression tree model, the preset regression tree model can predict the demand information based on the historical demand, the historical value attribute value and the historical browsing amount at the same time, so that the accuracy of the demand information prediction can be further improved.

For example, feature information composed of first feature information and second feature information corresponding to each article to be predicted may be subjected to a stitching process to obtain a feature matrix, such as And the characteristic matrix is input into the preset regression tree model, so that the preset regression tree model can predict the demand information of a plurality of articles at the same time, and the prediction efficiency is improved.

S450, detecting whether the current predicted date is the target predicted date, if not, executing the step S460, and if so, executing the step S490.

And S460, determining the estimated value attribute value of the target object in the current prediction date according to the historical demand and the historical value attribute value.

And S470, determining the estimated browsing amount of the target object in the current prediction date according to the historical demand and the historical browsing amount.

Specifically, the estimated browsing amount of the target object in the current prediction date is determined based on the historical demand and the historical browsing amount, so that demand information prediction of the subsequent prediction date can be performed based on the prediction demand and the estimated value attribute value corresponding to the current prediction date. When the predicted demand amount corresponding to the next prediction date is predicted, the predicted browsing amount corresponding to the current prediction date can be used as the historical browsing amount corresponding to the current prediction date to perform information prediction, so that the demand information prediction can be performed on the basis of the historical demand amount, the historical value attribute value and the historical browsing amount corresponding to the recent historical time period in each cycle, and the accuracy of demand information prediction is further improved.

Here, the execution sequence of step S470 is not limited. For example, step S470 may be performed sequentially after step S460, may be performed before step S460, or may be performed simultaneously with step S460.

Exemplarily, S470 may include: determining a first historical demand total amount of a target year in which the current forecast date is positioned and a second historical demand total amount of the target year in the same time period of the previous year according to the historical demand amounts; determining a demand increase proportion according to the first historical demand total amount and the second historical demand total amount; and determining the estimated browsing amount of the target item in the current prediction date according to the demand increase proportion and the historical browsing amount corresponding to the historical date in the historical time period which is in the same period as the current prediction date.

The historical date in the same period as the current prediction date in the historical time period may be a historical date in the same day as the current prediction date in the previous year of the target year. For example, if the current predicted date is 2021 year 11 month 11 day, the historical date in the same period as the current predicted date in the historical time period may be 2020 year 11 month 11 day.

Specifically, the historical demand amounts corresponding to each day before the current date in the target year of the current forecast date may be added to obtain the first historical demand total amount. And adding the historical demand quantities corresponding to each day in the same time period in the previous year of the target year to obtain a second historical demand total quantity. For example, if the current date is 2021 year 10 month 13 day, the daily demand amounts corresponding to each day between 2021 year 1 month 1 day and 2021 year 10 month 13 day may be added to obtain the first historical demand total amount. And adding the daily demand amounts corresponding to each day between 1 and 13 days 1 and 10 and 13 days 2020 to obtain a second historical demand total amount. And taking the ratio of the first historical demand sum to the second historical demand sum as the demand increase proportion. And multiplying the increase proportion of the demand quantity by the historical browsing quantity corresponding to the historical date in the same period as the current prediction date, and taking the obtained multiplication result as the estimated browsing quantity of the target object in the current prediction date, so that the flow of the future date can be accurately estimated.

S480, the operation of S410 is returned to the execution with the next date of the current predicted date as the current predicted date.

And S490, stopping the prediction operation.

According to the technical scheme, the first characteristic information and the second characteristic information are input into the preset regression tree model to predict the demand information, so that the preset regression tree model can predict the demand information based on the historical demand, the historical value attribute value and the historical browsing amount, and the accuracy of prediction of the demand information is further improved.

The following is an embodiment of an information prediction apparatus provided in an embodiment of the present invention, which belongs to the same inventive concept as the information prediction methods of the above embodiments, and reference may be made to the above embodiment of the information prediction method for details that are not described in detail in the embodiment of the information prediction apparatus.

Example four

Fig. 5 is a schematic structural diagram of an information prediction apparatus according to a fourth embodiment of the present invention, which is applicable to a case of predicting demand information of an article within any prediction date. As shown in fig. 5, the apparatus specifically includes: a historical data acquisition module 510, a first characteristic information determination module 520, an information prediction module 530, and a predictive value attribute value determination module 540.

The historical data acquiring module 510 is configured to acquire a historical demand and a historical value attribute value of the target item in a historical time period before a current prediction date, where an initial value of the current prediction date is the current date; a first characteristic information determining module 520, configured to determine, according to the historical demand and the historical value attribute value, first characteristic information corresponding to the target item; the information prediction module 530 is configured to input the first feature information into a preset regression tree model to perform demand information prediction, and determine a predicted demand of the target item within a current prediction date according to output of the preset regression tree model; and the estimated value attribute value determining module 540 is configured to determine an estimated value attribute value of the target item within the current prediction date according to the historical demand and the historical value attribute value, use the next date of the current prediction date as the current prediction date, and return to perform the operation of obtaining the historical demand and the historical value attribute value of the target item within the historical time period before the current prediction date until the current prediction date is the target prediction date.

Optionally, the predictive value attribute value determining module 540 includes:

the demand explosion coefficient determining unit is used for determining a demand explosion coefficient corresponding to the current prediction date according to the historical demand corresponding to each article belonging to the target category, wherein the target category is the category to which the target article belongs;

the month granularity regression coefficient acquisition unit is used for acquiring a month granularity regression coefficient corresponding to the target object;

the target discount proportion determining unit is used for determining a target discount proportion corresponding to a month previous month of a target month in which the current prediction date is based on a predetermined month granularity discount proportion corresponding to a target article;

the value attribute value discount proportion determining unit is used for determining the value attribute value discount proportion between a target month and a previous month according to the demand burst coefficient, the month granularity regression coefficient and the target discount proportion;

and the estimated value attribute value determining unit is used for determining the estimated value attribute value of the target item in the current prediction date according to the value attribute value discount proportion and the first value attribute value of the target item in the previous month.

Optionally, the monthly granularity regression coefficient obtaining unit is specifically configured to: determining a month granularity discount proportion corresponding to the target object according to the historical demand, the historical value attribute value and the historical object acquisition cost corresponding to each object belonging to the target category; determining the total monthly granularity article acquisition cost corresponding to the target article according to the historical demand and the historical article acquisition cost corresponding to each article belonging to the target category; and performing linear regression on the month granularity discount proportion and the total acquired cost of the month granularity articles, and determining a month granularity regression coefficient corresponding to the target article.

Optionally, the predictive value attribute value determining unit is specifically configured to: determining a value attribute value estimation coefficient corresponding to the target month according to the value attribute value discount proportion, the first value attribute value of the target item in the previous month and the second value attribute value of the target item in the target month in the historical time period; and determining the estimated value attribute value of the target object in the current prediction date according to the value attribute value estimation coefficient and the third value attribute value corresponding to the history date which is in the same period as the current prediction date in the history time period.

Optionally, the first feature information includes: demand statistics, value attribute value statistics, and cross-over features between demand and value attribute values.

Optionally, the first characteristic information determining module 520 is specifically configured to: determining a first variance according to the historical demand and determining a second variance according to the historical value attribute value; determining covariance between the demand and the value attribute value according to the historical demand and the historical value attribute value; and determining the cross feature between the demand and the value attribute value according to the first variance, the second variance and the covariance.

Optionally, the apparatus further comprises:

and the preset regression tree model training module is used for training the preset regression tree model according to the sample characteristic information and the actual demand based on a gradient descent mode before the preset regression tree model is used.

Optionally, the apparatus further comprises:

the second characteristic information determining module is used for acquiring the historical browsing amount of the target object in the historical time period before the current prediction date and determining second characteristic information corresponding to the target object according to the historical browsing amount;

the information prediction module 530 is specifically configured to: inputting the first characteristic information and the second characteristic information into a preset regression tree model for demand information prediction, and determining the predicted demand of the target object within the current prediction date according to the output of the preset regression tree model;

the device also includes: and the estimated browsing amount determining module is used for determining the estimated browsing amount of the target object in the current prediction date according to the historical demand and the historical browsing amount after determining the estimated value attribute value of the target object in the current prediction date.

Optionally, the second characteristic information includes: the statistical characteristics of the browsing amount and the cross characteristics between the demand amount and the browsing amount;

the second characteristic information determination module is specifically configured to: determining statistical characteristics of the browsing amount according to the historical browsing amount; and determining the cross characteristics between the demand and the browsing amount according to the historical browsing amount and the historical demand.

Optionally, the predicted browsing amount determining module is specifically configured to: determining a first historical demand total amount of a target year in which the current forecast date is positioned and a second historical demand total amount of the target year in the same time period of the previous year according to the historical demand amounts; determining a demand increase proportion according to the first historical demand total amount and the second historical demand total amount; and determining the estimated browsing amount of the target item in the current prediction date according to the demand increase proportion and the historical browsing amount corresponding to the historical date in the historical time period which is in the same period as the current prediction date.

Optionally, the apparatus further comprises: the target replenishment quantity determining module is used for acquiring the target forecast demand corresponding to each day in preset stocking days after the current date after determining the forecast demand of the target article in the target forecast date; determining a target inventory corresponding to the target object according to the target forecast demand, the corresponding numerical value of the preset spot rate level under normal distribution and the preset standard deviation; and determining the target replenishment quantity according to the target inventory quantity and the current inventory quantity.

The information prediction device provided by the embodiment of the invention can execute the information prediction method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of executing the information prediction method.

It should be noted that, in the embodiment of the information prediction apparatus, the included units and modules are merely divided according to functional logic, but are not limited to the above division as long as the corresponding functions can be implemented; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.

EXAMPLE five

Fig. 6 is a schematic structural diagram of an electronic device according to a fifth embodiment of the present invention. FIG. 6 illustrates a block diagram of an exemplary electronic device 12 suitable for use in implementing embodiments of the present invention. The electronic device 12 shown in fig. 6 is only an example and should not bring any limitation to the function and the scope of use of the embodiment of the present invention.

As shown in FIG. 6, electronic device 12 is embodied in the form of a general purpose computing device. The components of electronic device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that couples various system components including the system memory 28 and the processing unit 16.

Bus 18 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures. By way of example, such architectures include, but are not limited to, Industry Standard Architecture (ISA) bus, micro-channel architecture (MAC) bus, enhanced ISA bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.

Electronic device 12 typically includes a variety of computer system readable media. Such media may be any available media that is accessible by electronic device 12 and includes both volatile and nonvolatile media, removable and non-removable media.

The system memory 28 may include computer system readable media in the form of volatile memory, such as Random Access Memory (RAM)30 and/or cache memory 32. The electronic device 12 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be used to read from and write to non-removable, nonvolatile magnetic media (not shown in FIG. 6, and commonly referred to as a "hard drive"). Although not shown in FIG. 6, a magnetic disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk") and an optical disk drive for reading from or writing to a removable, nonvolatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 by one or more data media interfaces. System memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the invention.

A program/utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may comprise an implementation of a network environment. Program modules 42 generally carry out the functions and/or methodologies of the described embodiments of the invention.

Electronic device 12 may also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), with one or more devices that enable a user to interact with electronic device 12, and/or with any devices (e.g., network card, modem, etc.) that enable electronic device 12 to communicate with one or more other computing devices. Such communication may be through an input/output (I/O) interface 22. Also, the electronic device 12 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the Internet) via the network adapter 20. As shown, the network adapter 20 communicates with other modules of the electronic device 12 via the bus 18. It should be understood that although not shown in the figures, other hardware and/or software modules may be used in conjunction with electronic device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.

The processing unit 16 executes various functional applications and data processing by executing programs stored in the system memory 28, for example, to implement an information prediction method provided by the embodiment of the present invention, the method includes:

acquiring historical demand and historical value attribute values of a target article in a historical time period before a current prediction date, wherein an initial value of the current prediction date is the current date;

determining first characteristic information corresponding to the target object according to the historical demand and the historical value attribute value;

inputting the first characteristic information into a preset regression tree model for demand information prediction, and determining the predicted demand of the target object within the current prediction date according to the output of the preset regression tree model;

and determining the estimated value attribute value of the target item in the current prediction date according to the historical demand and the historical value attribute value, taking the next date of the current prediction date as the current prediction date, and returning to execute the operation of acquiring the historical demand and the historical value attribute value of the target item in the historical time period before the current prediction date until the current prediction date is the target prediction date.

Of course, those skilled in the art will understand that the processor may also implement the technical solution of the information prediction method provided in any embodiment of the present invention.

EXAMPLE six

The present embodiment provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the information prediction method steps as provided by any of the embodiments of the present invention, the method comprising:

acquiring historical demand and historical value attribute values of a target article in a historical time period before a current prediction date, wherein an initial value of the current prediction date is the current date;

determining first characteristic information corresponding to the target object according to the historical demand and the historical value attribute value;

inputting the first characteristic information into a preset regression tree model for demand information prediction, and determining the predicted demand of the target object within the current prediction date according to the output of the preset regression tree model;

and determining the estimated value attribute value of the target item in the current prediction date according to the historical demand and the historical value attribute value, taking the next date of the current prediction date as the current prediction date, and returning to execute the operation of acquiring the historical demand and the historical value attribute value of the target item in the historical time period before the current prediction date until the current prediction date is the target prediction date.

Computer storage media for embodiments of the invention may employ any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. The computer-readable storage medium may be, for example but not limited to: an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

A computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.

Computer program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).

It will be understood by those skilled in the art that the modules or steps of the invention described above may be implemented by a general purpose computing device, they may be centralized on a single computing device or distributed across a network of computing devices, and optionally they may be implemented by program code executable by a computing device, such that it may be stored in a memory device and executed by a computing device, or it may be separately fabricated into various integrated circuit modules, or it may be fabricated by fabricating a plurality of modules or steps thereof into a single integrated circuit module. Thus, the present invention is not limited to any specific combination of hardware and software.

It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.