Target object recommendation method and device, computer equipment and storage medium

文档序号：1963852 发布日期：2021-12-14 浏览：21次中文

阅读说明：本技术 一种目标对象推荐方法、装置、计算机设备及存储介质 (Target object recommendation method and device, computer equipment and storage medium ) 是由薛进葛生根于 2021-08-13 设计创作，主要内容包括：本发明公开了一种目标对象推荐方法、装置、计算机设备及存储介质,该方法包括：利用基于用户的协同过滤算法,根据目标用户的用户行为数据从候选用户群中筛选出与目标用户匹配的相似用户群,根据相似用户群对应的候选对象生成目标用户对应的第一候选对象列表,利用基于物品的协同过滤算法,根据目标用户的用户行为数据计算获取目标用户对应的第二候选对象列表,根据第一候选对象列表、第二候选对象列表以及预设规则计算获取确定推荐给目标用户的目标对象,通过利用基于用户的协同过滤算法获取与目标用户相似的用户喜欢的物品、利用基于物品的协同过滤算法获取与目标用户之前喜欢的物品相似的物品一起推荐给目标用户,提高推荐的准确度以及覆盖率。(The invention discloses a target object recommendation method, a target object recommendation device, computer equipment and a storage medium, wherein the method comprises the following steps: screening out similar user groups matched with the target user from the candidate user groups according to the user behavior data of the target user by utilizing a user-based collaborative filtering algorithm, generating a first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group, utilizing a collaborative filtering algorithm based on articles, calculating and acquiring a second candidate object list corresponding to the target user according to the user behavior data of the target user, calculating and obtaining a target object recommended to a target user according to the first candidate object list, the second candidate object list and a preset rule, the method and the device improve the accuracy and coverage rate of recommendation by acquiring the articles similar to the target user and liked by the user through the user-based collaborative filtering algorithm, and acquiring the articles similar to the articles previously liked by the target user and recommending the articles to the target user through the article-based collaborative filtering algorithm.)

1. A target object recommendation method, characterized in that the method comprises:

screening out a similar user group matched with a target user from candidate user groups according to user behavior data of the target user by using a user-based collaborative filtering algorithm;

generating a first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group;

calculating and acquiring a second candidate object list corresponding to the target user according to the user behavior data of the target user by using a collaborative filtering algorithm based on articles;

and calculating and obtaining a target object which is determined to be recommended to the target user according to the first candidate object list, the second candidate object list and a preset rule.

2. The method for recommending a target object according to claim 1, wherein the step of screening out a similar user group matching the target user from the candidate user group according to the user behavior data of the target user by using a user-based collaborative filtering algorithm comprises:

calculating user behavior data of a target user and user behavior data corresponding to candidate user groups by adopting a preset similarity algorithm to obtain the similarity between the target user and each candidate user group;

and determining the candidate user group corresponding to the similarity meeting the preset threshold as a similar user group matched with the target user.

3. The method of claim 2, further comprising an optimization process of the preset similarity algorithm, comprising:

and analyzing the calculation factors of the preset similarity algorithm, and combining the calculation factors meeting the combination requirement and/or deleting the calculation factors not meeting the calculation requirement according to the analysis result.

4. The method of any one of claims 1 to 3, wherein the generating the first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group comprises:

and acquiring all candidate objects corresponding to the similar user group, and screening all candidate objects according to a preset screening rule to generate a first candidate object list corresponding to the target user.

5. The method for recommending a target object according to any one of claims 1 to 3, wherein the calculating and acquiring a second candidate object list corresponding to the target user according to the user behavior data of the target user by using an article-based collaborative filtering algorithm comprises:

determining the current user behavior of the target user according to the user behavior data of the target user, acquiring a corresponding candidate object according to the current user behavior, and generating a second candidate object list corresponding to the target user.

6. The method for recommending a target object according to claim 5, wherein the determining the current user behavior of the target user according to the user behavior data of the target user and the obtaining a corresponding candidate object according to the current user behavior comprises:

determining the current user behavior of the target user according to the user behavior data of the target user, and acquiring a first object corresponding to the current user behavior;

and clustering the first object, and determining a candidate object corresponding to the current behavior of the user according to a clustering result.

7. The method of claim 6, wherein the determining the current user behavior of the target user according to the user behavior data of the target user, and the obtaining the corresponding candidate object according to the current user behavior further comprises:

determining the user intention of the target user according to the current user behavior, acquiring a second object corresponding to the user intention, and determining a candidate object corresponding to the current user behavior according to the clustering result and the second object.

8. A target object recommendation apparatus, the apparatus comprising:

the user screening module is used for screening out a similar user group matched with a target user from a candidate user group according to user behavior data of the target user by utilizing a user-based collaborative filtering algorithm;

the first calculation module is used for generating a first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group;

the second calculation module is used for calculating and acquiring a second candidate object list corresponding to the target user according to the user behavior data of the target user by utilizing an article-based collaborative filtering algorithm;

and the object determining module is used for calculating and obtaining a target object recommended to the target user according to the first candidate object list, the second candidate object list and a preset rule.

9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the method of any of claims 1 to 7 are implemented when the computer program is executed by the processor.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.

Technical Field

The present invention relates to the field of data processing technologies, and in particular, to a target object recommendation method and apparatus, a computer device, and a storage medium.

Background

With the rapid increase in the number of user accesses over the internet, the amount of information created and accessed by users has increased exponentially. Therefore, many internet enterprises hope to record as much data as possible of users by using a multi-dimensional and multi-channel user behavior data acquisition mode in the process of using products by the users, and on the other hand, the enterprises are difficult to accurately analyze the data of the users aiming at the large data of the users in such a scale, so that the optimal service cannot be effectively put into the most suitable user crowd. In the prior art, the recommendation mode adopted for recommending products based on user data is generally single, and the problems of low recommendation accuracy and coverage rate and the like exist.

Therefore, it is desirable to provide a target object recommendation method to solve the above problems.

Disclosure of Invention

In order to solve the problems in the prior art, embodiments of the present invention provide a target object recommendation method, an apparatus, a computer device, and a storage medium, so as to overcome the problems in the prior art that a recommendation manner adopted for product recommendation based on user data is generally single, and recommendation accuracy and coverage are not high.

In order to solve one or more technical problems, the invention adopts the technical scheme that:

in a first aspect, a target object recommendation method is provided, which includes the following steps:

screening out a similar user group matched with a target user from candidate user groups according to user behavior data of the target user by using a user-based collaborative filtering algorithm;

generating a first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group;

Further, the screening out a similar user group matched with the target user from the candidate user group according to the user behavior data of the target user by using the user-based collaborative filtering algorithm includes:

and determining the candidate user group corresponding to the similarity meeting the preset threshold as a similar user group matched with the target user.

Further, the method further includes an optimization process of the preset similarity algorithm, including:

Further, the generating a first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group includes:

Further, the calculating and acquiring a second candidate object list corresponding to the target user according to the user behavior data of the target user by using the article-based collaborative filtering algorithm includes:

Further, the determining the current user behavior of the target user according to the user behavior data of the target user, and acquiring the corresponding candidate object according to the current user behavior includes:

determining the current user behavior of the target user according to the user behavior data of the target user, and acquiring a first object corresponding to the current user behavior;

and clustering the first object, and determining a candidate object corresponding to the current behavior of the user according to a clustering result.

Further, the determining the current user behavior of the target user according to the user behavior data of the target user, and obtaining the corresponding candidate object according to the current user behavior further includes:

In a second aspect, there is provided a target object recommendation apparatus, the apparatus comprising:

the first calculation module is used for generating a first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group;

In a third aspect, a computer device is provided, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the computer program, the following steps are implemented:

screening out a similar user group matched with a target user from candidate user groups according to user behavior data of the target user by using a user-based collaborative filtering algorithm;

generating a first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group;

In a fourth aspect, there is provided a computer readable storage medium having a computer program stored thereon, which when executed by a processor, performs the steps of:

screening out a similar user group matched with a target user from candidate user groups according to user behavior data of the target user by using a user-based collaborative filtering algorithm;

generating a first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group;

The technical scheme provided by the embodiment of the invention has the following beneficial effects:

in the target object recommendation method, apparatus, computer device, and storage medium provided in embodiments of the present invention, a user-based collaborative filtering algorithm is used to screen a similar user group matching a target user from candidate user groups according to user behavior data of the target user, a first candidate object list corresponding to the target user is generated according to candidate objects corresponding to the similar user group, an article-based collaborative filtering algorithm is used to calculate and obtain a second candidate object list corresponding to the target user according to user behavior data of the target user, a target object recommended to the target user is obtained and determined according to the first candidate object list, the second candidate object list, and a preset rule, and an article liked by a user similar to the target user is obtained by using the user-based collaborative filtering algorithm, and acquiring articles similar to articles liked by the target user before by using an article-based collaborative filtering algorithm, recommending the articles to the target user, and improving the accuracy and coverage rate of recommendation.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

FIG. 1 is a flow diagram illustrating a target object recommendation method in accordance with an exemplary embodiment;

FIG. 2 is a schematic diagram illustrating the structure of a target object recommendation apparatus in accordance with an exemplary embodiment;

FIG. 3 is a schematic diagram illustrating an internal architecture of a computer device, according to an example embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

As described in the background, as the number of user accesses on the Internet has increased rapidly, the amount of information created and accessed by users has increased exponentially. Therefore, many internet enterprises hope to record as much data as possible of users by using a multi-dimensional and multi-channel user behavior data acquisition mode in the process of using products by the users, and on the other hand, the enterprises are difficult to accurately analyze the data of the users aiming at the large data of the users in such a scale, so that the optimal service cannot be effectively put into the most suitable user crowd. Taking commodity recommendation as an example, in the prior art, the recommendation mode adopted for commodity recommendation based on user data is generally single, and the problems of low recommendation accuracy and coverage rate and the like exist.

In order to solve the above problems, the embodiment of the present invention creatively provides a target object recommendation method, which reduces unnecessary exposure waste by analyzing potential users and delivering commodities to specific user groups, and constructs a user-oriented personalized recommendation system by analyzing relevance of user data, so as to implement customized deployment of thousands of people and thousands of faces for services or products. In specific implementation, on one hand, a collaborative filtering algorithm based on the user is adopted to recommend the user to the articles similar to the interests of the user and on the other hand, a collaborative filtering algorithm based on the articles is adopted to recommend the user to the articles similar to the articles previously liked by the user, so that the recommendation accuracy and the recommendation coverage rate are improved.

Fig. 1 is a flowchart illustrating a target object recommendation method according to an exemplary embodiment, and referring to fig. 1, the method includes the following steps:

s1: and screening out a similar user group matched with the target user from the candidate user group according to the user behavior data of the target user by utilizing a user-based collaborative filtering algorithm.

In particular, the simplest form of presence of user behavior data is a log. In the embodiment of the application, when user behavior data is obtained, the original logs may be collected into session logs (session logs) according to user behaviors, where each session represents one-time user behavior and corresponding services, such as an exposure log, a click log, an additional purchase log, a payment log, and the like. User behaviors are generally classified into explicit feedback behaviors (explicit feedback) and implicit feedback behaviors (implicit feedback) in a personalized recommendation system. The explicit feedback behavior includes behavior in which a user explicitly indicates a preference for an article, and the implicit feedback behavior refers to behavior in which the preference of the user cannot be explicitly reflected.

Specifically, for the convenience of subsequent calculation, in the embodiment of the present application, a user portrait including, but not limited to, an individual user portrait and a group user portrait may be constructed in advance for information generated by user behaviors under a big data condition. In specific implementation, effective structure of explicit knowledge is taken as a target, extraction of multi-source user information is achieved by adopting an entity recognition technology, a relation extraction technology and an attribute extraction technology based on deep semantic learning, information fusion of users across modes and fields is achieved by adopting a knowledge fusion technology based on a cross-mode shared subspace learning theory, dynamic evolution and updating of a knowledge map and the like are achieved by adopting a knowledge inference and entity expansion technology based on a deep neural language model, and details are not repeated here.

Specifically, in the embodiment of the present application, the user behavior analysis includes analyzing a distribution relationship between user activity and popularity of an item, where the item is a target object. The user behavior data meets the long tail distribution (power law), and analysis according to a preset formula can find that the more active the user is, the more the user tends to be for cold goods.

S2: and generating a first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group.

Specifically, in the embodiment of the application, a collaborative filtering algorithm based on the user is adopted to recommend articles liked by other users similar to the interests of the target user. In specific implementation, a user group with the same interest as the target user (i.e. a similar user group matched with the target user) is found, then a target object liked by the users in the user group is found, and the target object is recommended to the target user.

S3: and calculating and acquiring a second candidate object list corresponding to the target user according to the user behavior data of the target user by using a collaborative filtering algorithm based on articles.

Specifically, in the embodiment of the present application, an article-based collaborative filtering algorithm is adopted to recommend an article similar to an article that the target user previously likes. As a better implementation manner, in the embodiment of the present application, a pre-constructed user interest model may be used to recommend a target object. In specific implementation, the user behavior data set is uniformly and randomly divided into M parts, preferably, M is 8, one part is selected as a test set, and the rest is selected as a training set. And establishing a user interest model on the training set, predicting the user behavior on the test set, and counting corresponding evaluation indexes. In order to ensure that the evaluation index is not an overfitting result, M tests are needed, different test sets are used each time, and then the average value of the evaluation indexes of the M tests is taken as a final evaluation index.

In order to prevent the result of a certain experiment from being an overfitting result (overfitting), different numbers of users and the same random seed can be selected each time, M different training sets and test sets can be obtained by performing M experiments, then the experiments are performed respectively, and the average value of the M experiments is used as the final evaluation index.

Specifically, in the embodiment of the present application, the evaluation index includes, but is not limited to, the following indexes:

1. accuracy or recall

Recall is used to describe how many proportions of user-item scoring records are included in the final recommendation list, and accuracy describes how many proportions of user-item scoring records in the final recommendation list have occurred. Assuming that N items are recommended to a user u and are marked as (R (u)), the item set liked by the user on the test set is marked as (T (u)), and then the precision of the recommendation algorithm is evaluated through precision (precision) or recall (recall), and the formula is as follows:

2. coverage rate

The coverage rate reflects the capability of the recommendation algorithm to explore the long tail, the higher the coverage rate is, the more the recommendation algorithm can recommend the articles in the long tail to the user, and the formula is as follows:

3. degree of novelty

And measuring the novelty of the recommendation result by using the average popularity of the items in the recommendation list, if the recommended items are popular, indicating that the recommended novelty is low, otherwise indicating that the recommendation result is novel.

S4: and calculating and obtaining a target object which is determined to be recommended to the target user according to the first candidate object list, the second candidate object list and a preset rule.

Specifically, after obtaining an article (i.e., a first candidate object list) liked by another user and similar to the interest of the target user based on the collaborative filtering algorithm of the user and obtaining an article (i.e., a second candidate object list) liked by an article before the target user based on the collaborative filtering algorithm of the article, the objects in the first candidate object list and the second candidate object list may be scored according to a preset rule, ranked according to the scores, and then the target objects meeting preset requirements are screened and recommended to the target user, for example, the objects ranked in the top several names are screened and recommended to the target user as the target objects, and the like, which is not limited herein, the user may set the object according to actual requirements.

As a preferred implementation manner, in this embodiment of the present application, the screening, by using a user-based collaborative filtering algorithm, a similar user group matching a target user from a candidate user group according to user behavior data of the target user includes:

and determining the candidate user group corresponding to the similarity meeting the preset threshold as a similar user group matched with the target user.

Specifically, in the embodiment of the application, the interest similarity of two users is calculated mainly by using the similarity of user behaviors. The preset similarity algorithm includes, but is not limited to, a jaccard similarity algorithm or a cosine similarity algorithm, etc.

Assuming that two users with similarity to be calculated are u and v respectively, let N (u) be an article set that user u has positive feedback, and let N (v) be an article set that user v has positive feedback, calculating the similarity between the two users by a jaccard similarity algorithm as follows:

the similarity formula between the two users is calculated by a cosine similarity algorithm as follows:

as a preferred implementation manner, in an embodiment of the present application, the method further includes an optimization process of the preset similarity algorithm, including:

Specifically, there are many calculation factors for calculating the similarity, including but not limited to the following:

proximity: the method is used for describing the influence of the scoring difference between the users on the similarity;

significance: for describing the relationship between the user score and the median of the scoring domain, which can be used to distinguish whether the user likes or dislikes the item;

singularity: the method is used for describing the influence of the difference value of the average value of the scores of two similar users on a common score item and the global average value of the score item on the similarity;

jaccard factor: the method is used for considering the influence of the number of the common scoring items of the users on the similarity;

urp (user Rating preference): for considering the influence of the difference in the mean of the scores between users and the difference in the variance of the scores of users on the similarity.

On one hand, some of the above factors require additional calculation, which makes the calculation process of the similarity algorithm very complicated, and each calculation link may bring errors, and the superposition of multiple calculation errors may increase the possibility of deviation from the actual value, and may cause the occurrence of a zero division phenomenon in the product form. On the other hand, the scoring granularity of most scoring data sets is not large enough, and the use of the calculation factor does not necessarily increase the discrimination between user groups, but increases the calculation complexity and the calculation error. In order to solve the above problems, in the embodiment of the present application, each calculation factor of the preset similarity algorithm is analyzed, and then the calculation factors are correspondingly processed according to the analysis result, including but not limited to merging the calculation factors meeting the merging requirement, deleting the calculation factors not meeting the calculation requirement, increasing the threshold value judgment, and judging the conditions of unreasonable calculation factors and increasing new calculation factors according to the conditions of unreasonable calculation factors.

As a preferred implementation manner, in this embodiment of the present application, the generating a first candidate object list corresponding to the target user according to the candidate object corresponding to the similar user group includes:

Specifically, for example, in the case of commodity recommendation, after obtaining articles liked by other users similar to the interest of a target user based on a collaborative filtering algorithm of the user, in order to improve the accuracy of recommendation, implement commodity delivery for a specific user group, and reduce unnecessary exposure waste, in the embodiment of the present application, all obtained candidate objects may be screened according to a preset screening rule, commodities already purchased by the target user in the candidate objects are removed, and a first candidate object list corresponding to the target user is generated for the remaining candidate objects.

As a better implementation manner, in this embodiment of the application, the calculating and obtaining, by using a collaborative filtering algorithm based on an article, a second candidate object list corresponding to the target user according to the user behavior data of the target user includes:

Specifically, in the embodiment of the application, when a target object is recommended for a target user, for example, the recommended coverage rate is improved, user behavior data can be analyzed, the current behavior of the target user is determined, the requirement of the target user is captured, and then a corresponding business is recommended to the target user.

As a better implementation manner, in this application embodiment, the determining the current user behavior of the target user according to the user behavior data of the target user, and acquiring a corresponding candidate object according to the current user behavior includes:

determining the current user behavior of the target user according to the user behavior data of the target user, and acquiring a first object corresponding to the current user behavior;

and clustering the first object, and determining a candidate object corresponding to the current behavior of the user according to a clustering result.

Specifically, data such as the current commodity browsing duration and the click behavior amount in the user behavior data are identified, the strong interest of the user in certain commodities is determined according to the data such as the current commodity browsing duration and the click behavior amount, then the commodities are clustered, the clustering result is used as a candidate object recommended to the target user, for example, if the target user is identified to browse certain commodity currently, the commodity and similar commodities obtained by clustering the commodity are recommended to the target user. Therefore, the path for the user to browse the commodities can be shortened, and the conversion rate is improved.

Specifically, after the current user behavior of the target user is determined, the user visiting intention (i.e., the user intention) can be predicted through a combination of multiple behavior modes, including but not limited to a search behavior, a collection behavior, a purchase adding behavior, and the like of the user, so as to locate whether the user has a strong demand for purchasing goods today or to view historical purchased goods logistics information. For example, if the user has recently searched for "masks" but has not made a transaction, then a recommendation may be made to the user for products such as masks.

As a preferred implementation, in the present application example, when the recommendation policy is made, long-term preference and real-time preference of the user may be combined, where the user profile refers to long-term preference of the user, for example, long-term preference of a user is women's dress, but since a user has just recently handed in to boyfriends, short-term preference of today is men's dress. Because the preference of the user decays along with time, the real-time preference data can represent the current demand of the user, and the recommendation effect is more accurate. The real-time preference can be generally inferred through searching, navigation screening or user browsing of the user, if the user does not form a closed transaction loop, such as browsing, buying, and unpaid, we can assume that his needs are not met, possibly that a proper goods is not found, and possibly that other friend purchases are made because of a high price, and at this time, we recommend the goods preferred by the user in real time to the user, and the conversion rate is often better.

In addition to this, people do not like complete strangeness and always want to be able to find out the elements that are familiar with from new things. Thus a balance point can be found between novelty and familiarity when recommending a target object for a user.

From human perspective: counting the repetition degree of the categories which are frequently purchased by the user in the past, if the repetition rate is low, indicating that the user relatively prefers novel commodities, increasing the ratio of the commodities which are not purchased at this time, and vice versa;

starting from a scene: daily users may like to purchase goods purchased historically as well as regular related items, and during short periods or on special festive days users may spread a wider range of items for purchase.

The existing recommendation logic is single in form, not enough in interestingness and too few in recommendation dimensionality, so that the dimensionality constructed by a recommendation strategy needs to be expanded, and the recommendation dimensionality is complementary to the dimensionality constructed by the user hierarchy and the content morphology, so that a better effect is achieved. In specific implementation, on one hand, the users can be divided according to various dimensions, for example, the users are divided into user groups with medium and high customer unit prices and user groups with low customer unit prices according to user consumption capacity, or the users are divided into student parties, moms, families with cars and the like, identity recognition is obtained through identity tags, and meanwhile, the user requirements of long tails can be met. On the other hand, the recommendation may be performed according to the content organization of the object to be recommended, such as directly recommending activity-based preferential commodities, recommending explosive money in forms of ranking lists, thermal-marketing lists, and the like, and recommending interesting small commodities.

Fig. 2 is a schematic structural diagram illustrating a target object recommending apparatus according to an exemplary embodiment, the apparatus including:

the first calculation module is used for generating a first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group;

As a preferred implementation manner, in an embodiment of the present application, the user filtering module is configured to:

and determining the candidate user group corresponding to the similarity meeting the preset threshold as a similar user group matched with the target user.

As a preferred implementation manner, in the embodiment of the present application, the apparatus further includes:

and the algorithm optimization module is used for analyzing the calculation factors of the preset similarity algorithm, and combining the calculation factors meeting the combination requirement and/or deleting the calculation factors not meeting the calculation requirement according to the analysis result.

As a better implementation manner, in this embodiment of the application, the first computing module is configured to obtain all candidate objects corresponding to the similar user group, and filter all candidate objects according to a preset filtering rule to generate a first candidate object list corresponding to the target user.

As a better implementation manner, in this embodiment of the application, the second computing module is configured to determine the user current behavior of the target user according to the user behavior data of the target user, obtain a corresponding candidate object according to the user current behavior, and generate a second candidate object list corresponding to the target user.

As a better implementation manner, in this embodiment of the application, the second computing module is configured to determine a user current behavior of the target user according to the user behavior data of the target user, and obtain a first object corresponding to the user current behavior; and clustering the first object, and determining a candidate object corresponding to the current behavior of the user according to a clustering result.

As a preferred implementation manner, in this embodiment of the application, the second calculating module is configured to determine a user intention of the target user according to the current behavior of the user, obtain a second object corresponding to the user intention, and determine a candidate object corresponding to the current behavior of the user according to the clustering result and the second object.

Fig. 3 is a schematic diagram illustrating an internal configuration of a computer device according to an exemplary embodiment, which includes a processor, a memory, and a network interface connected through a system bus, as shown in fig. 3. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a method of optimization of an execution plan.

Those skilled in the art will appreciate that the configuration shown in fig. 3 is a block diagram of only a portion of the configuration associated with aspects of the present invention and is not intended to limit the computing devices to which aspects of the present invention may be applied, and that a particular computing device may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

As a preferred implementation manner, in an embodiment of the present invention, the computer device includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor implements the following steps when executing the computer program:

screening out a similar user group matched with a target user from candidate user groups according to user behavior data of the target user by using a user-based collaborative filtering algorithm;

generating a first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group;

As a preferred implementation manner, in the embodiment of the present invention, when the processor executes the computer program, the following steps are further implemented:

and determining the candidate user group corresponding to the similarity meeting the preset threshold as a similar user group matched with the target user.

As a preferred implementation manner, in the embodiment of the present invention, when the processor executes the computer program, the following steps are further implemented:

determining the current user behavior of the target user according to the user behavior data of the target user, and acquiring a first object corresponding to the current user behavior;

and clustering the first object, and determining a candidate object corresponding to the current behavior of the user according to a clustering result.

As a preferred implementation manner, in the embodiment of the present invention, when the processor executes the computer program, the following steps are further implemented:

In an embodiment of the present invention, a computer-readable storage medium is further provided, on which a computer program is stored, and when the computer program is executed by a processor, the computer program implements the following steps:

screening out a similar user group matched with a target user from candidate user groups according to user behavior data of the target user by using a user-based collaborative filtering algorithm;

generating a first candidate object list corresponding to the target user according to the candidate objects corresponding to the similar user group;

As a preferred implementation manner, in the embodiment of the present invention, when executed by the processor, the computer program further implements the following steps:

and determining the candidate user group corresponding to the similarity meeting the preset threshold as a similar user group matched with the target user.

As a preferred implementation manner, in the embodiment of the present invention, when executed by the processor, the computer program further implements the following steps:

determining the current user behavior of the target user according to the user behavior data of the target user, and acquiring a first object corresponding to the current user behavior;

and clustering the first object, and determining a candidate object corresponding to the current behavior of the user according to a clustering result.

As a preferred implementation manner, in the embodiment of the present invention, when executed by the processor, the computer program further implements the following steps:

In summary, the technical solution provided by the embodiment of the present invention has the following beneficial effects:

It should be noted that: in the target object recommendation device provided in the foregoing embodiment, when triggering a recommendation service, only the division of the functional modules is illustrated, and in practical applications, the function distribution may be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules, so as to complete all or part of the functions described above. In addition, the target object recommendation apparatus provided in the above embodiment and the target object recommendation method embodiment belong to the same concept, that is, the apparatus is based on the target object recommendation method, and the specific implementation process thereof is detailed in the method embodiment and is not described herein again.

It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

16页详细技术资料下载

Target object recommendation method and device, computer equipment and storage medium

相关技术

网友询问留言