Land parcel selection method and device, electronic equipment and storage medium

文档序号：1875579 发布日期：2021-11-23 浏览：15次中文

阅读说明：本技术 地块选择方法、装置、电子设备及存储介质 (Land parcel selection method and device, electronic equipment and storage medium ) 是由吴非权孙福宁杨帆许迅腾王文来张升王诏远于 2021-09-09 设计创作，主要内容包括：本申请实施例提供了一种地块选择方法、装置、电子设备及存储介质,涉及人工智能技术领域,可以用于地图、车联网等场景。该方法包括：获取到与地块选择相关的样本对象数据,基于预设的地块画像数据与样本对象数据,确定与样本对象所在地块相应的候选地块；基于预设的地块画像数据对候选地块进行相似度检索,以在候选地块中确定与样本对象所在地块相应的目标地块；其中,地块画像数据包括基于地理空间划分所得地块相应的画像数据。本申请方案的实施可以有效提高地块选择的精度。(The embodiment of the application provides a method and a device for selecting a region, electronic equipment and a storage medium, relates to the technical field of artificial intelligence, and can be used in scenes such as maps, internet of vehicles and the like. The method comprises the following steps: acquiring sample object data related to plot selection, and determining candidate plots corresponding to plots where sample objects are located based on preset plot image data and sample object data; performing similarity retrieval on the candidate plots based on preset plot image data so as to determine target plots corresponding to the plots where the sample objects are located in the candidate plots; the image data of the land parcel comprises image data corresponding to the land parcel obtained by dividing based on the geographic space. The implementation of the scheme of the application can effectively improve the accuracy of land parcel selection.)

1. A method for selecting a block, comprising:

acquiring sample object data related to plot selection, and determining candidate plots corresponding to plots where sample objects are located based on preset plot image data and the sample object data;

performing similarity retrieval on the candidate plots based on preset plot image data so as to determine target plots corresponding to the plots where the sample objects are located in the candidate plots;

the image data of the land parcel comprises image data corresponding to the land parcel obtained by dividing based on the geographic space.

2. The method of claim 1, wherein determining the candidate parcel corresponding to the parcel where the sample object is based on the preset parcel image data and the sample object data comprises:

performing automatic training of a model based on preset plot image data and the sample object data to obtain a target inference model;

determining a candidate plot corresponding to the plot where the sample object is located based on the target inference model;

the target inference model is an inference model obtained by selecting at least one model from multiple fused models in a semi-supervised automatic learning and training mode.

3. The method of claim 2, wherein the performing of automatic model training based on the preset parcel image data and the sample object data to obtain a target inference model comprises:

carrying out automatic feature selection on preset plot image data based on the sample object data to obtain plot sample features of each plot;

carrying out automatic training on a model based on training data determined by the corresponding relation between the plot where the sample object is located and the plot sample characteristics to obtain a first inference model;

reasoning the characteristics of the parcel sample without the parcel based on the first reasoning model, and combining the reasoning result data with the confidence coefficient higher than a preset threshold value with the training data to obtain combined training data; the land sample characteristics of the unconfigured land comprise land characteristics which do not have corresponding relation with the land where the sample object is located;

carrying out automatic training by adopting the combined training data to obtain a target inference model;

wherein, the automatic training comprises adopting a classification model to evaluate indexes to automatically adjust model parameters.

4. The method according to claim 3, wherein the automatically training the model based on the training data determined by the corresponding relationship between the plot of the sample object and the sample feature of the plot to obtain the first inference model comprises:

determining initial training data based on the corresponding relation between the plot where the sample object is located and the plot sample characteristics;

sampling the initial training data to obtain processed training data;

and carrying out automatic training of the model based on the processed training data to obtain a first inference model.

5. The method of claim 3, wherein the automatically training with the merged training data to obtain the target inference model comprises:

carrying out automatic training by adopting the combined training data to obtain a second reasoning model;

based on the received user-defined feature information and the user-defined weight coefficient, adjusting the combined training data;

and carrying out automatic training by adopting the adjusted training data to obtain a target inference model.

6. The method of claim 5, wherein adjusting the combined training data based on the received custom feature information and custom weight coefficients comprises:

extracting at least one training characteristic information adopted for training the second reasoning model and a training weight coefficient thereof to be displayed on a user interface of the client;

and adjusting the training characteristic information and the training weight coefficient thereof based on the received user-defined characteristic information and the user-defined weight coefficient to obtain adjusted training data.

7. The method of claim 1, wherein the performing similarity search on the candidate parcel based on preset parcel image data to determine a target parcel corresponding to the parcel where the sample object is located in the candidate parcel comprises:

acquiring basic features constructed based on preset plot portrait data;

extracting candidate features of the candidate land parcels;

and performing feature similarity retrieval on the basic features and the candidate features, and determining a target plot corresponding to the plot where the sample object is located.

8. The method of claim 7, wherein constructing the base features based on the pre-defined parcel image data comprises:

determining a characteristic value of a preset index based on preset plot image data;

performing down-sampling processing on the characteristic value to obtain a characteristic value after dimensionality reduction;

and analyzing the feature importance based on the feature value after dimension reduction to obtain the basic feature.

9. The method of claim 8, wherein determining a feature value of a predetermined indicator based on predetermined parcel image data comprises at least one of:

under a set scale, determining a characteristic value of an adjacent entropy based on the type number of sample objects in an adjacent range of a preset radius of a place where a parcel is located;

determining the characteristic value of the competitive degree based on the number of the same types of sample objects in the adjacent range of the preset radius of the position of the land parcel under the set scale;

under a set scale, determining a feature value of a plot space correlation effect based on the set type number of the sample objects in the adjacent range of the preset radius of the position of the plot and an attraction coefficient between the type of the sample objects;

under a set scale, determining a characteristic value of the transfer quality based on the visited number of users in the proximity range of the preset radius of the position of the parcel and the transfer probability between the types of the sample objects.

10. The method according to claim 7, wherein the performing feature similarity search on the basic features and the candidate features to determine a target region corresponding to a region where the sample object is located comprises:

performing feature similarity retrieval on the basic features and the candidate features to determine a first land block in the candidate land blocks;

and determining a target plot corresponding to the plot where the sample object is located in the first plot based on at least one of the transaction times of the point of interest POI and the custom selection information.

11. The method of claim 1,

before acquiring sample object data related to the selection of the parcel, the method further comprises the following steps:

responding to a plot selection request operation of a client;

after the target land parcel is determined, the method further comprises the following steps:

displaying plot selection information corresponding to a target plot at the client;

the displaying of the parcel selection information corresponding to the target parcel comprises: and displaying the position of the target plot on the map interface based on a preset mark form.

12. A block selection apparatus, comprising:

the system comprises a first determination module, a second determination module and a third determination module, wherein the first determination module is used for acquiring sample object data related to plot selection and determining candidate plots corresponding to plots where sample objects are located based on preset plot image data and the sample object data;

the second determination module is used for carrying out similarity retrieval on the candidate land parcels based on preset land parcel image data so as to determine a target land parcel corresponding to the land parcel where the sample object is located in the candidate land parcels;

the image data of the land parcel comprises image data corresponding to the land parcel obtained by dividing based on the geographic space.

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

one or more processors;

a memory;

one or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs configured to: performing the method according to any one of claims 1 to 11.

14. A computer-readable storage medium for storing computer instructions which, when executed on a computer, cause the computer to perform the method of any of claims 1 to 11.

15. A computer program product comprising a computer program or instructions for implementing the steps of the method of any one of claims 1 to 11 when executed by a processor.

Technical Field

The present application relates to the field of artificial intelligence technologies, and in particular, to a method and an apparatus for selecting a region, an electronic device, a computer-readable storage medium, and a computer program product.

Background

The land parcel selection refers to a process of selecting an address before building. The factors to be considered for the selection of the plot are various, such as population density, traffic, existing buildings and the like, so that basic data with huge data volume need to be considered during the selection of the plot; in addition, the plot selection needs to be considered in combination with the development requirements of the city or the enterprise, that is, the demand data with less data volume needs to be considered when the plot is selected.

Disclosure of Invention

The technical scheme provided by the application aims to solve at least one of the technical defects, particularly the technical defect of low accuracy of land parcel selection. The technical scheme is as follows:

in a first aspect of the present application, there is provided a method of selecting a block, comprising:

the image data of the land parcel comprises image data corresponding to the land parcel obtained by dividing based on the geographic space.

In an embodiment, the determining, based on preset plot image data and the sample object data, a candidate plot corresponding to a plot where the sample object is located includes:

performing automatic training of a model based on preset plot image data and the sample object data to obtain a target inference model;

determining a candidate plot corresponding to the plot where the sample object is located based on the target inference model;

the target inference model is an inference model obtained by selecting at least one model from multiple fused models in a semi-supervised automatic learning and training mode.

In an embodiment, the performing automatic training of a model based on preset parcel portrait data and the sample object data to obtain a target inference model includes:

carrying out automatic feature selection on preset plot image data based on the sample object data to obtain plot sample features of each plot;

carrying out automatic training by adopting the combined training data to obtain a target inference model;

wherein, the automatic training comprises adopting a classification model to evaluate indexes to automatically adjust model parameters.

In an embodiment, the performing automatic training of the model based on the training data determined by the correspondence between the parcel in which the sample object is located and the parcel sample feature to obtain the first inference model includes:

determining initial training data based on the corresponding relation between the plot where the sample object is located and the plot sample characteristics;

sampling the initial training data to obtain processed training data;

and carrying out automatic training of the model based on the processed training data to obtain a first inference model.

In an embodiment, the automatically training the merged training data to obtain the target inference model includes:

carrying out automatic training by adopting the combined training data to obtain a second reasoning model;

based on the received user-defined feature information and the user-defined weight coefficient, adjusting the combined training data;

and carrying out automatic training by adopting the adjusted training data to obtain a target inference model.

In an embodiment, the adjusting the combined training data based on the received custom feature information and the custom weight coefficient includes:

In an embodiment, the performing similarity retrieval on the candidate parcel based on preset parcel image data to determine a target parcel corresponding to a parcel where the sample object is located in the candidate parcel comprises:

acquiring basic features constructed based on preset plot portrait data;

extracting candidate features of the candidate land parcels;

and performing feature similarity retrieval on the basic features and the candidate features, and determining a target plot corresponding to the plot where the sample object is located.

In one embodiment, constructing the base features based on the preset plot image data includes:

determining a characteristic value of a preset index based on preset plot image data;

performing down-sampling processing on the characteristic value to obtain a characteristic value after dimensionality reduction;

and analyzing the feature importance based on the feature value after dimension reduction to obtain the basic feature.

In an embodiment, the determining the feature value of the preset index based on the preset plot image data includes at least one of the following:

under a set scale, determining a characteristic value of an adjacent entropy based on the type number of sample objects in an adjacent range of a preset radius of a place where a parcel is located;

In an embodiment, the performing feature similarity retrieval on the basic features and the candidate features to determine a target region corresponding to a region where the sample object is located includes:

performing feature similarity retrieval on the basic features and the candidate features to determine a first land block in the candidate land blocks;

In one embodiment, before obtaining sample object data related to the selection of the parcel, the method further comprises:

responding to a plot selection request operation of a client;

after the target land parcel is determined, the method further comprises the following steps:

displaying plot selection information corresponding to a target plot at the client;

the displaying of the parcel selection information corresponding to the target parcel comprises: and displaying the position of the target plot on the map interface based on a preset mark form.

In a second aspect of the present application, there is provided a block selection apparatus including:

the image data of the land parcel comprises image data corresponding to the land parcel obtained by dividing based on the geographic space.

In a third aspect of the present application, there is provided an electronic device including:

one or more processors;

a memory;

one or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs configured to: the method provided by the first aspect is performed.

In a fourth aspect of the present application, a computer-readable storage medium is provided for storing computer instructions which, when executed on a computer, cause the computer to perform the method provided by the first aspect.

In a fifth aspect of the present application, a computer program product is provided, comprising a computer program or instructions, characterized in that the computer program or instructions, when executed by a processor, implements the steps of the method provided in the first aspect.

The beneficial effect that technical scheme that this application provided brought is:

in the application, when sample object data related to plot selection is acquired, firstly, candidate plots corresponding to plots where sample objects are located are determined based on preset plot image data and sample object data, and the number of the obtained candidate plots is large; on the basis, similarity retrieval is carried out on the candidate plots based on preset plot image data, the plot selection result is further reduced, and the accuracy of plot selection is improved, so that the target plots corresponding to the plots where the sample objects are located are determined in the candidate plots. The image data of the land parcel comprises image data corresponding to the land parcel obtained by dividing based on the geographic space. The implementation of the scheme of the application can effectively mine the existing plot image data and the sample object data provided by the user, is beneficial to improving the accuracy of plot selection, and enables the selected plot to be more appropriate to the user requirements.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings used in the description of the embodiments of the present application will be briefly described below.

Fig. 1 is a flowchart of a method for selecting a block according to an embodiment of the present application;

fig. 2 is a system flow diagram of a method for selecting a parcel according to an embodiment of the present application;

FIG. 3 is a block diagram illustrating a process of model automated training in a method for selecting a parcel according to an embodiment of the present disclosure;

FIG. 4 is a block diagram illustrating a process of model automated training in a method for selecting a parcel according to an embodiment of the present disclosure;

FIG. 5 is a block diagram illustrating a process of building a basic feature based on a predetermined parcel portrait data in a parcel selection method according to an embodiment of the present disclosure;

fig. 6 is a block diagram illustrating a flow of similarity retrieval in a method for selecting a region according to an embodiment of the present disclosure;

FIG. 7 is a block flow diagram of a culled parcel in a parcel selection method according to an embodiment of the present application;

fig. 8 is a system flow diagram of another parcel selection method according to an embodiment of the present application;

fig. 9 is a schematic view of an interaction environment applied to a parcel selection method according to an embodiment of the present application;

fig. 10 is a schematic structural diagram of a parcel selection apparatus according to an embodiment of the present application;

fig. 11 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are exemplary only for the purpose of explaining the present application and are not to be construed as limiting the present application.

As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. It will be understood that when an element is referred to as being "connected" or "coupled" to another element, it can be directly connected or coupled to the other element or intervening elements may also be present. Further, "connected" or "coupled" as used herein may include wirelessly connected or wirelessly coupled. As used herein, the term "and/or" includes all or any element and all combinations of one or more of the associated listed items.

To make the objects, technical solutions and advantages of the present application more clear, embodiments of the present application will be described in further detail below with reference to the accompanying drawings.

The following is a description of the technology and nomenclature involved in this application:

AI (Artificial Intelligence) is a theory, method, technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and expand human Intelligence, perceive the environment, acquire knowledge and use the knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technique of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a manner similar to human intelligence. Artificial intelligence is the research of the design principle and the realization method of various intelligent machines, so that the machines have the functions of perception, reasoning and decision making. The artificial intelligence technology is a comprehensive subject and relates to the field of extensive technology, namely the technology of a hardware level and the technology of a software level. The artificial intelligence infrastructure generally includes technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and the like.

In the present application, machine learning/deep learning, etc. directions are possible.

The ML (Machine Learning) is a multi-domain cross subject, and relates to multiple subjects such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory, and the like. The special research on how a computer simulates or realizes the learning behavior of human beings so as to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve the performance of the computer. Machine learning is the core of artificial intelligence, is the fundamental approach for computers to have intelligence, and is applied to all fields of artificial intelligence. Machine learning and deep learning generally include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and formal education learning.

In the method for selecting a parcel provided by the embodiment of the application, a machine learning correlation technique may be used to process a small amount of sample object data, and a deep learning correlation technique may be used to process a large amount of parcel data.

However, in the related art, since data mining cannot be performed while accommodating basic data with a large amount of data and demand data with a small amount of data, various data characteristics cannot be fully utilized, and the reasonability of the finally selected parcel is very poor. In addition, in the related art, the method for embedding the user knowledge mainly directly specifies the weight for linear calculation, but for massive data, the similar simple processing cannot fully utilize the feature expression capability of the mined data, and the data are difficult to be used in a scene.

To solve at least one of the above problems, the present application provides a method, an apparatus, an electronic device, and a computer-readable storage medium for selecting a location; the method can be used for maps, internet of vehicles and other scenes. Specifically, sample object data provided by a user and existing plot portrait data can be fully utilized, automatic machine learning automatic ML plot selection analysis of mass data is supported, and experience of the user can be fused. The application provides an algorithm frame fitting scene design, and the rationality of land parcel selection can be effectively improved.

The following describes the technical solutions of the present application and how to solve the above technical problems with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.

In the embodiment of the present application, a method for selecting a parcel is provided, as shown in fig. 1, fig. 1 shows a schematic flow diagram of the method for selecting a parcel provided in the embodiment of the present application, where the method may be executed by any electronic device, such as a user terminal, or a server, where the user terminal may be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, a vehicle-mounted device, and the like, the server may be an independent physical server, or a server cluster or distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as a cloud service, a cloud database, a cloud computing, a cloud function, a cloud storage, a Network service, cloud communication, a middleware service, a domain name service, a security service, a CDN (Content Delivery Network), and a big data and artificial intelligence platform, the present application is not so limited.

The following first describes relevant contents of the parcel selection method provided in the embodiment of the present application.

By taking the example of setting an offline store in a selected plot of an enterprise, and assuming that the selected plot is in a plot selection scene for setting a store, the plot selection method provided by the embodiment of the application can realize that a plot suitable for setting the store is selected by the sample store based on sample store data provided by a user. The plot can be used for certain application purposes, such as planning, plot selection and the like, within a certain geographic spatial distribution range; with respect to shape, a parcel may include a regular parcel and an irregular parcel. In the embodiment of the present application, a certain geographic spatial distribution range (e.g., a certain administrative region) may be divided into rectangular land blocks with a set length, and the land blocks obtained by the division serve as a basis for the implementation of the present application; however, the embodiment of the present application may also be applied to other land parcel planning methods, and the embodiment of the present application does not limit this.

Specifically, the following steps S101 to S102 included in the method are described with reference to fig. 1, fig. 2, and fig. 8:

step S101: sample object data related to the selection of the plot is acquired, and a candidate plot corresponding to the plot where the sample object is located is determined based on preset plot image data and the sample object data.

In a field selected by the plot, a user may upload known relevant data of some sample objects to the plot selection platform, and the plot selection platform may fuse the data provided by the user with existing plot data to perform processing, and screen out candidate plots corresponding to the plots where the sample objects are located (that is, find out candidate plots with higher relevance based on the correspondence between the sample objects and the plots where the sample objects are located).

The sample object data may include attribute information of the sample object itself, and may include, for example, when the sample object is a clothing store, operation data of the clothing store, an area of the store, a location where the store is located (which may correspond to a location of a parcel), and the like. Alternatively, the sample object data that can be acquired in step S101 is stored in units of sample objects, that is, one sample object corresponds to one piece of sample object data, and how much sample object data depends on data that can be provided by the user.

The plot portrait data is data for analyzing service by abstracting each concrete information of the plot into a label based on big data and utilizing the label to concretize the plot image. The portrait data of the plot 1 may include the geographic location, population information within a certain range centered on the geographic location, traffic information, etc.; the feature that the land image data can include is various. Specifically, the image data of the land comprises image data corresponding to the land divided based on the geographic space.

Comparing the data size of the sample object data with that of the land image data, the sample object data belongs to the data to be analyzed with a small data size, and the land image data belongs to the existing data with a huge data size.

Specifically, the implementation of the above step S101 can be realized by a machine learning related technology, and a specific implementation process will be described in the following embodiments.

Step S102: and performing similarity retrieval on the candidate plots based on preset plot image data so as to determine a target plot corresponding to the plot where the sample object is located in the candidate plots.

In the embodiment of the present application, on the basis that the selected result of the plot obtained by screening based on step S101 is rough, that is, the candidate plot may include many plots, and the precision of the plot selection is not high, the present application adopts step S102 to perform the similarity matching filtering step based on the plot image data, and performs similarity retrieval on the candidate plots, so as to further reduce the number of plots included in the selected result of the plot, and improve the rationality and precision of the plot selection.

The similarity search may be a feature search performed by using a vector search algorithm, which is used to further optimize the candidate blocks, that is, the number of the candidate blocks is greater than the number of the target blocks. Alternatively, the target parcel may comprise at least one parcel.

Specifically, the implementation of the step S102 can be realized by a deep learning related technology, and a specific implementation process will be described in the following embodiments.

In a possible embodiment, the method for selecting a parcel provided by the embodiment of the present application may further include only step S101, that is, outputting the candidate parcel as a parcel selection result; it is also possible to include only step S102, that is, to perform the parcel selection process directly based on parcel image data without fusing sample object data provided by the user.

A specific procedure for determining a candidate parcel corresponding to a parcel in which a sample object is located is described below.

In one embodiment, the step S101 of determining a candidate parcel corresponding to the parcel where the sample object is located based on the preset parcel image data and the sample object data includes steps a1-a 2:

step A1: and carrying out automatic training on the model based on preset plot portrait data and the sample object data to obtain a target inference model.

In the embodiment of the application, it is considered that in some plot selection scenes, after a user submits sample object data, a plot selection algorithm needs to automatically give a plot selection result, but automatic learning and reasoning are not supported in the related technology, and the plot selection algorithm in the related technology is difficult to copy in batch, so that the problem of difficult operation and maintenance exists (for example, when sample objects are catering shops and clothing shops, learned plot selection algorithms are different, and if the sample objects need to be processed for different types, the calculation complexity is very high, and the operation and maintenance cost is very high). Therefore, the embodiment of the application adopts automatic machine learning (AutoML) to solve the problem.

The automatic machine learning AutoML is an end-to-end process automation process for applying machine learning to real problems, and a machine learning model and an AI model are trained in a motorized training mode. Specifically, as shown in fig. 3 and 4, AutoML can realize automation from three aspects of feature engineering, model construction and parameter optimization.

Specifically, preset parcel portrait data and sample object data can be used for building training data (feature engineering) of automatic learning training of the inference model, then a proper model is selected from a frame structure fused with a plurality of models through automatic training for building, automatic parameter adjustment is finally achieved, and the building of the target inference model is completed.

Step A2: and determining a candidate land parcel corresponding to the land parcel where the sample object is located based on the target inference model.

Specifically, the target inference model is a model trained based on sample object data currently provided by a user and existing parcel image data, and a relatively suitable candidate parcel can be selected from the existing parcels (known through parcel image data) through the target inference model based on the sample object data to serve as an output of the model. If the sample object data provided by the user is the data of the catering stores, other candidate plots which are relatively suitable for establishing the catering stores can be screened out through the target inference model.

In the embodiment of the application, the target inference model may be an inference model constructed by selecting at least one model from multiple fused models through a semi-supervised automatic learning training mode. The following description is directed to an automatic learning training process of the target inference model.

In an embodiment, in the step a1, based on the preset parcel portrait data and the sample object data, performing an automatic training of a model to obtain a target inference model, including the steps a11-a 14:

step A11: and carrying out automatic feature selection on preset plot portrait data based on the sample object data to obtain plot sample features of each plot.

Specifically, step a11 belongs to the execution step of automatic feature engineering in AutoML. In the machine learning step, the cost of labor and time consumed by the feature engineering is high, the automatic feature engineering can be used for automatic operation, and the model training efficiency is improved.

The characteristics are abstract numerical transformation from a concrete object to numerical representation, and the characteristic engineering refers to a process of converting original data into characteristics, so that the characteristics can be better expressed to the model, and the accuracy of the model for data processing is improved.

Alternatively, the processing of automatic feature engineering may include feature completion, normalization, outlier processing, and the like.

In the automatic feature selection process, considering that thousands of hundreds of features can be received based on the plot portrait data, the number of the features is far larger than that of the sample objects, not all the features are beneficial to machine understanding, and not all the features are related to modeling content, some features with higher importance can be selected and are suitable for the plot sample features required for generating the construction model of the application. Alternatively, if the sample object data includes the traffic of the store a, the data having a high correlation with the traffic of the store a may be selected from the plot drawing data to generate the plot sample feature. In some embodiments, automated feature selection may also not be performed based on sample object data, i.e., feature selection may be performed only on the terrain imagery data; for example, the plot image data can be applied to various scenes, and feature selection can be performed according to the application angle of plot selection by adapting to the plot selection method of the application.

Optionally, after the automatic feature selection, the feature types in the feature of the parcel sample corresponding to each parcel may be the same or different; which can be set according to actual requirements.

Step A12: and carrying out automatic training on the model based on training data determined by the corresponding relation between the plot where the sample object is located and the plot sample characteristics to obtain a first inference model.

Specifically, the region where the sample object is located can be determined based on the sample object data provided by the user, and then the corresponding relationship between the region where the sample object is located and the characteristics of the region sample can be determined. If the sample object A is located in the plot B, the corresponding relation is formed between the plot sample characteristics C, D and E; it will be appreciated that with respect to the parcel sample features C, D and E, there is a configured tag parcel B, which has a mapping relationship with the sample object A.

The corresponding relation between the plot where the sample object is located and the plot sample characteristics can construct training data of the positive sample, and further automatic training of the model can be carried out based on the training data of the positive sample, so that the first inference model is obtained. That is, the first inference model is trained based on positive sample training data constructed from sample object data provided by the user.

In an embodiment, in step a12, based on training data determined by the correspondence between the parcel in which the sample object is located and the parcel sample characteristics, performing automatic training of the model to obtain a first inference model, includes the following steps a121-a 123:

step A121: determining initial training data based on the corresponding relation between the plot where the sample object is located and the plot sample characteristics;

step A122: sampling the initial training data to obtain processed training data;

step A123: and carrying out automatic training of the model based on the processed training data to obtain a first inference model.

Specifically, considering that the amount of sample object data provided by the user is small, in order to further improve the training data proportion of the positive sample, a training data enhancement step may be performed, and the model is trained automatically by using the enhanced training data. The training data enhancement may adopt a technique such as Borderline-SMOT (boundary line synthesis minority class oversampling) to perform oversampling processing, which is a refinement method mainly oriented to unbalanced data sets (e.g., in the present application, the sample object data is less, and the data amount of the feature of the block sample is larger).

Alternatively, the sampling process performed on the initial training data may be up-sampling or down-sampling. The data volume of the training data can be improved through the up-sampling processing; the down-sampling process can extract the training data, and further improve the proportion of positive samples in the training data.

Step A13: and reasoning the characteristics of the parcel samples without the parcels based on the first reasoning model, and combining the reasoning result data with the confidence coefficient higher than a preset threshold value with the training data to obtain combined training data.

Specifically, because the data amount of the sample object data provided by the user is small, there may exist a large number of parcel sample features that are not configured with parcels, that is, training data without tags, and for the training data without tags, the embodiment of the present application may adopt the first inference model to perform inference, that is, the training data without tags is used as input data of the first inference model, and the first inference model infers the parcels corresponding to the relevant features.

In consideration of the low inference accuracy of the first inference model, in the execution of the subsequent steps, inference result data with a confidence higher than a preset threshold may be selected to be merged with training data (data for training the first inference model constructed based on the sample object data) to obtain merged training data.

The land sample features of the unconfigured land comprise land features which do not have corresponding relation with the land where the sample object is located.

Step A14: and carrying out automatic training by adopting the combined training data to obtain a target inference model.

Specifically, the combined training data may be used to continue training the first inference model, or the initialized model may be retrained, so as to obtain the final target inference model.

Optionally, the AutoML automatic training provided in the embodiment of the present application belongs to a process of model construction, and may be a framework of multi-model fusion, which makes full use of characteristics of various data, and selects one or more suitable models to construct an inference model required by the present application. It can be understood that the algorithms used to construct the inference model may be different to accommodate different plot selection requirements. For example, the target inference model may be a clustering model, a multi-classification model, or the like.

Wherein, the automatic training comprises adopting a classification model to evaluate indexes to automatically adjust model parameters.

Specifically, the automatic training process includes automatic adjustment of model parameters, and a greedy search method based on a KS value can be adopted to perform automatic parameter adjustment in the embodiment of the present application. The KS value is an evaluation index used for distinguishing and predicting the separation degree of positive and negative samples in the model, the prediction result of each sample is probability or a fraction range, and the cumulative distribution of the positive and negative samples is from the minimum probability or the minimum fraction to the maximum probability or the maximum fraction; the KS value is the absolute value of the largest difference in the two distributions. The value range of the KS value is [0,1], and the larger the KS value is, the better the positive and negative sample distinguishing degree is. Wherein greedy search is to select the output value with the highest probability in each step to obtain a decoded output sequence (i.e. to provide a search path corresponding to a search space); it is understood that a label obtained through model inference can correspond to a plurality of paths, and therefore, the path with the highest probability is not equal to the final label with the highest probability value. Specifically, when the automatic parameter adjustment is performed by the greedy search method based on the KS value, only a single parameter is considered each time to perform forward and backward search so as to obtain the optimal parameter.

In the embodiment of the present application, as shown in fig. 3 and 4, a model training method of semi-supervised automatic learning is specifically adopted through the implementation of steps a11-a 14.

In a possible embodiment, the step a14 uses the merged training data to perform automatic training to obtain the target inference model, which includes steps a141-a 143:

step A141: carrying out automatic training by adopting the combined training data to obtain a second reasoning model;

step A142: based on the received user-defined feature information and the user-defined weight coefficient, adjusting the combined training data;

step A143: and carrying out automatic training by adopting the adjusted training data to obtain a target inference model.

As shown in fig. 3, after the second inference model is obtained by performing automatic training based on the combined training data, feature information (which may be a feature type) and a weight coefficient thereof used in the training process of the second inference model may be extracted and displayed to the user; if the user determines that the feature type and the corresponding weight coefficient need to be adjusted according to the self requirement, the user-defined feature information and the corresponding weight coefficient can be input, and the processing can be the addition and the filtration of the relevant features in the sample plot feature set. That is, in the embodiment of the application, the user can dynamically add or delete the characteristics of the sample plot according to the actual experience, so that the contact degree between the plot selection result and the user's own requirements can be effectively improved, and the user experience is improved.

Specifically, after the second inference model is obtained through training, at least one training feature information adopted for training the second inference model and a training weight coefficient corresponding to the training feature information can be extracted and displayed on a user interface of the client; the user can consider whether the data of the training target inference model needs to be adjusted according to the plot selection requirement through the displayed training characteristic information and the corresponding training weight coefficient so as to obtain a plot selection result more suitable for the self requirement. Furthermore, the user can directly adjust the training characteristic information and the training weight coefficient thereof, and can also input new characteristic information and the corresponding weight coefficient thereof by himself, and data generated by the user in the operation process of the user interface can be regarded as corresponding self-defined characteristic information and self-defined weight coefficient; on this basis, adjusted training data may be obtained.

A specific procedure for determining a target parcel corresponding to a parcel where the sample object is located is described below.

In one embodiment, the step S102 of performing similarity search on the candidate parcel based on preset parcel image data to determine a target parcel corresponding to the parcel where the sample object is located in the candidate parcel comprises steps B1-B3:

step B1: and acquiring basic features constructed based on preset land parcel portrait data.

Specifically, as shown in fig. 5, the existing land image data in the embodiment of the present application may be constructed based on the basic data, and under the basic data, the corresponding land image data is constructed for each divided land. In addition, in consideration of adaptability of the plot image data to the plot selection method provided in the embodiment of the present application, in the construction of the feature data, in addition to the image features of the traffic, the point of interest (poi), the population, and the like of the plot, it is necessary to construct features for the plot selection. Optionally, the feature data may include dimension-based proximity entropy, competitiveness, jensen quality (different plot space-related contributions), population transfer density, transfer quality, and tenant feature, among others. According to the method and the device, a set of index system can be set for practical application of land parcel selection, and characteristic values corresponding to various indexes are determined; it is understood that step B1 is an execution step of feature extraction and preprocessing for the parcel image data.

In the geographic information system, a POI may refer to a point having physical significance, such as a house, a shop, a mailbox, a bus station, and the like.

In one embodiment, building the base features based on the pre-defined tile image data includes steps C1-C3:

step C1: and determining a characteristic value of a preset index based on preset land parcel portrait data.

Specifically, the method and the device can set an index system from the practical application angle selected from the plot, and consider the index of the related type by combining with the scale factor. If the sample object is a store, the type may be a clothing store, a restaurant, a toy store, or the like.

In one possible embodiment, the determining the feature value of the preset index in step C1 based on the preset parcel image data includes at least one of the following steps C11-C14:

step C11: under a set scale, determining a characteristic value of the proximity entropy based on the number of types of sample objects in a proximity range of a preset radius of a place where the parcel is located.

In particular, the neighboring entropy x based on scale_l(r) can be calculated using the following equation (1):

wherein l represents a land position, and r represents a preset radius; nr (l, r) represents the number of types in the vicinity of the l position corresponding to the r radius; it will be appreciated that on different plot scales, different fineness classification systems are employed. ST refers to all types of all specified dimensions.

Based on the above equation (1), it can be seen that the larger the proximity entropy, the larger the type diversity of the region.

Step C12: and under a set scale, determining the characteristic value of the competitive degree based on the number of the same types of the sample objects in the adjacent range of the preset radius of the position of the land parcel.

In particular, the scale-based degree of competition x_l(r) can be calculated using the following equation (2):

wherein l represents a land position, and r represents a preset radius; n is a radical of_Srl(l, r) refers to the same type number of a certain radius r at a particular scale. Here also the types are associated with scales, different scales having corresponding taxonomies.

Step C13: under a set scale, determining a feature value of the plot space correlation effect based on the set type number of the sample objects in the proximity range of the preset radius of the position of the plot and the attraction coefficient between the type of the sample objects.

In particular, the dimension-based Jensen mass x_l(r) (plot space correlation) can be calculated using the following equation (3);

wherein l represents a land position, and r represents a preset radius;refers to how many types r are on average at a specified scale, at a specified position of type rl_pIs observed. ST refers to a set of types that specify a scale. k is a radical of_rp→ rl refers to the attraction coefficient between types.

Step C14: under a set scale, determining a characteristic value of the transfer quality based on the visited number of users in the proximity range of the preset radius of the position of the parcel and the transfer probability between the types of the sample objects.

In particular, a scale-based transfer quality index x_l(r) can be calculated using the following equation (4):

wherein l represents a land position, and r represents a preset radius; c_pThe number of visits by the user at the designated position P is specified; sigma_Srp→SrlRefers to the transition probabilities of two types of Srp and Srl at a certain scale.

In particular, the index may be used to measure the user's attractiveness of a given location at a scale to other locations within the area. Such as assigning a user attraction of location 1 (e.g., restaurant B) to other locations 2 and 3 (e.g., restaurants C and D) within the area at a certain scale a.

Step C2: and performing down-sampling processing on the characteristic value to obtain the characteristic value after dimensionality reduction.

The data size of the features related to the land parcel is very large, so that the data dimension reduction can be performed on the features and the data sparsity is reduced.

Step C3: and analyzing the feature importance based on the feature value after dimension reduction to obtain the basic feature.

Specifically, the execution of step C3 may be understood as a processing step of feature importance analysis and feature screening, and the feature importance analysis may be performed by using algorithms such as xgboost (eXtreme Gradient Boosting) and linear regression. In addition, the method can also be used for carrying out weighting analysis on the feature weight by combining with a geographic detector, and screening the features to obtain the basic features.

Optionally, step C3 also involves feature deletion, where multiple rounds of feature deletion are performed based on the model applied to the importance analysis, the process may be performed by monitoring KS index values and PSI index values of the model, and if the index values do not change (have small differences) or tend to change well, the corresponding features may be deleted. The KS index value can be determined by adopting the KS value, the PSI index value is a population stability index, and the smaller the PSI value is, the smaller the difference between two distributions is, the more stable the distribution is.

Alternatively, as shown in fig. 6, the processing of steps C1-C3 may correspond to the execution process of feature engineering and feature Embedding extraction.

In a possible embodiment, the base features applied in the similarity matching filtering based on the plot data may be different from or the same as the plot sample features applied in the automatic learning training for a small amount of sample object data.

Step B2: and extracting candidate features of the candidate land blocks.

Specifically, the candidate features of the candidate land parcel may be extracted based on the feature engineering and feature Embedding as in fig. 6, or may be obtained by creating an index relationship between the basic features and the land parcel directly on the basis of the basic features. Through index establishment, the search of the features can be rapidly carried out, and the efficiency of subsequent feature retrieval is improved.

Step B3: and performing feature similarity retrieval on the basic features and the candidate features, and determining a target plot corresponding to the plot where the sample object is located.

Specifically, similarity retrieval of the candidate land blocks can be performed based on deep learning and a vector engine, wherein the similarity retrieval of the candidate land blocks is realized by performing feature similarity retrieval on the basic features and the candidate features. The implementation of step B3 is to select a parcel from the candidate parcels based on the basic features of the parcel image data, and obtain a target parcel corresponding to the parcel where the sample object is located.

As shown in fig. 6, after the feature retrieval is performed, approximate parcel data corresponding to relevant data of the target parcel may be acquired.

In one embodiment, in step B3, performing feature similarity search on the basic features and the candidate features, and determining a target parcel corresponding to a parcel where the sample object is located, includes steps B31-B32:

step B31: and performing feature similarity retrieval on the basic features and the candidate features to determine a first land block in the candidate land blocks.

Step B32: and determining a target plot corresponding to the plot where the sample object is located in the first plot based on at least one of the transaction times of the point of interest POI and the custom selection information.

Specifically, as shown in fig. 7, in order to further improve the rationality of the parcel selection and the degree of contact with the user requirement, the embodiment of the present application may further filter the parcel based on the first parcel obtained by performing similar matching filtering on the candidate parcel, and then based on the number of POI transactions, the user-defined selection information, and other data, to obtain the fine-filtered selected parcel (target parcel).

Alternatively, as shown in fig. 7, a model loading spark calculation engine, spark being a fast and general calculation engine designed for large-scale data processing, may be applied to perform various operations, and in the embodiment of the present application, loading spark may be applied to machine learning, that is, the process of determining the first parcel (that is, may correspond to the target parcel of step B3) through model loading spark completing step B31.

The self-defined selection information can be set based on the personalized plot selection requirements of the user, for example, plot selection information such as "the distance from a subway station cannot be more than 300 meters", "no competitor shop can exist within 100 meters" and the like, that is, the first plot can be screened by combining with geospatial analysis according to the user requirements, so that a selected plot (a final target plot) is obtained.

In a possible embodiment, before acquiring the sample object data related to the selection of the parcel in step S101, the method further includes: and responding to the land parcel selection request operation of the client.

Specifically, in the embodiment of the present application, the sample object data may be acquired in response to a parcel selection request operation initiated by a user through a client.

Optionally, the user may provide a requirement for selecting a parcel while initiating a parcel selection request operation, such as an administrative region (geographic space) for defining parcel selection, an object type for parcel selection (for example, a specific type of a certain enterprise when the shop is an entity to be opened), and the like.

After the target parcel is determined in step S102, the method further includes: and displaying the plot selection information corresponding to the target plot at the client.

Wherein displaying the parcel selection information corresponding to the target parcel comprises: and displaying the position of the target plot on the map interface based on a preset mark form.

Specifically, the preset mark form may include displaying a position indication tag at the position of the target parcel, and displaying longitude and latitude information, parcel size, parcel seating direction, and the like on the map interface corresponding to the position of the target parcel.

Specifically, after the target parcel corresponding to the parcel where the sample object is located is determined, parcel selection information corresponding to the target parcel may be displayed on a user interface of the client. The land selection information may include longitude and latitude coordinates of the land, the size of the land, the direction of the land, and feature information used for selecting the land. The display mode of the parcel selection information may be a mode of displaying the position of the target parcel on the map in the form of the map by a specific label or effect. The data format transmitted to the front end of the client may adopt a json (JavaScript Object notification) format.

Optionally, in this embodiment of the present application, the target parcel includes at least one parcel, and when there are more than one parcel, the target parcel may be sorted from near to far based on the current geographic location of the user, and finally, each piece of parcel selection information after sorting is displayed.

A feasible application example is provided below with reference to fig. 8 and 9 for the land parcel selection method provided in the embodiment of the present application.

Take the example that the user is enterprise a and wants to set up a movie theater as an entity in administrative area B.

The user can upload data (sample object data) related to a movie theater through the terminal 100, wherein the data related to the movie theater is not limited to only data defining the movie theater itself, but may also include data of other objects within the proximity of the movie theater, such as data related to restaurants, toy stores, etc. provided in a partition of the movie theater. Although the user wishes to select a land parcel in the administrative region B, the administrative region where the sample object is located in the sample object data uploaded by the user is not limited; because even sample object data in other administrative areas have a certain reference value for selecting a parcel in administrative area B.

Based on the plot selection requirement of the user (which may be the plot selection requirement submitted by the user interface triggering the plot selection request operation of the terminal 100 after the user uploads the data related to the movie theater), the existing plot image data may be obtained for plot selection. In the embodiment of the present application, considering that a user wishes to select a parcel in an administrative area B and the quantity of parcel image data is huge, the parcel image data corresponding to the administrative area B may be acquired only to perform parcel selection processing.

As shown in fig. 8, after receiving data about a movie theater uploaded by a user, characteristics of a parcel sample can be obtained (obtained in a characteristic engineering module), and training data is constructed based on sample object data uploaded by the user and the characteristics of the parcel sample; in addition, considering that the data volume of the sample object data uploaded by the user is small and the data volume of the corresponding training data is also small, the constructed training data can be subjected to data enhancement processing (oversampling processing) so as to perform automatic ml automatic training based on the enhanced training data, thereby obtaining the first inference model. Further, the embodiment of the application adopts a semi-supervised automatic learning training mode, so that a first reasoning model can be adopted to reason about the characteristics of the parcel sample without the parcel, and then reasoning result data with confidence higher than a preset threshold is merged with training data (which can be enhanced training data) to obtain merged training data; and then, carrying out automatic training by adopting the combined training data to obtain a second reasoning model. At this time, the feature type and the corresponding weight used for constructing the second inference model can be extracted and fed back to the user at the user interface of the client, and if the user considers that the feature type and the weight coefficient need to be adjusted in combination with the plot selection requirement of the user, the user-defined feature type and the user-defined weight coefficient can be input, so that the user-trimmed training data is further used for carrying out the automatic ML automatic training again, and the target inference model is finally obtained. The target inference model is an inference model obtained by automatically performing model training by a framework of a parcel selection algorithm after a user submits data related to a movie theater. At this point, in the large framework of the plot selection algorithm, the target inference model may determine candidate plots based on the movie theater related data uploaded by the user. And determining the candidate region blocks based on the region blocks where the sample objects uploaded by the user are located. The sample object may include one or more; at least one candidate parcel may be determined to accommodate a parcel in which a sample object is located.

As shown in FIG. 8, the candidate plots are output as a model, which is the basis data for the final determination of the culled plots. In the similarity matching filtering based on the plot data, the processed basic data is also the output of the target inference model, namely the candidate plot. It will be appreciated that a first parcel is determined by making a parcel selection among the candidate parcels based on the parcel image data.

The input data of the similarity matching filtering based on the plot data can be feature data extracted based on the plot portrait data, and can also be various feature values or basic features corresponding to preset indexes. In the processing of similarity matching filtering based on the plot data, firstly, feature extraction (including feature engineering between feature extractions) and preprocessing (such as index establishment) are carried out on the plot portrait data; similarity retrieval may then be performed for the candidate parcel based on the deep learning and vector engine. Wherein the base feature and the candidate feature can be processed into a base feature vector and a candidate feature vector. After performing the feature retrieval, approximate parcel data (data associated with the first parcel) may be obtained. It will be appreciated that the number of first plots is less than the number of candidate plots, and that the first plots may include at least one.

As shown in fig. 8, the relevant data of the first parcel can be input into the model loaded with spark, and the machine learning model can filter only the first parcel to obtain a rough parcel, and can also filter in combination with the relevant data of the candidate parcel to obtain a rough parcel; in addition, after the rough parcel is obtained, fine filtering can be performed based on the POI transaction frequency grade and/or the user-defined selection information, and finally a fine parcel (target parcel) is obtained. The user-defined selection information can be input when the plot selection request is initiated, or can be input at any time point in the plot selection process. Specifically, the custom selection information may be "no other movie theaters can exist within the same business circle", "no distance from a subway station or a bus station can exceed 100 meters", "no-work-day traffic is more than three times that of a work-day" or the like.

After the target parcel is finally determined, parcel selection information corresponding to the target parcel may be displayed on a user interface of the client. Such as displaying the geographic location (in administrative area B) of the target parcel on a star on the map interface.

As shown in fig. 9, the parcel selection method provided in the embodiment of the present application may be executed on the terminal 100, or may be executed on the server 200.

When the tile selection method is performed on the terminal 100, the preset tile image data may be stored in the server 200 or a database corresponding to the server 200, and the terminal 100 may acquire the tile image data from the server 200 through the network 300 when the tile selection method is performed.

When the tile selection method is executed on the server 200, after the user uploads the sample object data to the server 200 through the network 300 at the terminal 100, the server 200 executes the tile selection method to determine the target tile and then feeds back the target tile to the terminal 100 through the network 300, and finally, tile selection information corresponding to the target tile is displayed on the user interface of the terminal 100.

An embodiment of the present application provides a parcel selection apparatus, as shown in fig. 10, the parcel selection apparatus 900 may include: a first determining module 101 and a second determining module 102.

The first determining module 101 is configured to acquire sample object data related to parcel selection, and determine a candidate parcel corresponding to a parcel where a sample object is located based on preset parcel image data and the sample object data; the second determining module 102 is configured to perform similarity retrieval on the candidate parcel based on preset parcel image data to determine a target parcel corresponding to a parcel where the sample object is located in the candidate parcel; the image data of the land parcel comprises image data corresponding to the land parcel obtained by dividing based on the geographic space.

In an embodiment, the first determining module 101, when configured to determine a candidate parcel corresponding to a parcel where the sample object is located based on preset parcel image data and the sample object data, is specifically configured to:

performing automatic training of a model based on preset plot image data and the sample object data to obtain a target inference model;

determining a candidate plot corresponding to the plot where the sample object is located based on the target inference model;

the target inference model is an inference model obtained by selecting at least one model from multiple fused models in a semi-supervised automatic learning and training mode.

In an embodiment, the first determining module 101 is specifically configured to, when the first determining module is configured to perform automatic training of a model based on preset parcel portrait data and the sample object data to obtain a target inference model:

carrying out automatic feature selection on preset plot image data based on the sample object data to obtain plot sample features of each plot;

carrying out automatic training by adopting the combined training data to obtain a target inference model;

wherein, the automatic training comprises adopting a classification model to evaluate indexes to automatically adjust model parameters.

In an embodiment, the first determining module 101 is specifically configured to, when the first determining module is configured to perform automatic training of a model based on training data determined by a correspondence between a parcel in which a sample object is located and a feature of a sample of the parcel, to obtain a first inference model:

determining initial training data based on the corresponding relation between the plot where the sample object is located and the plot sample characteristics;

sampling the initial training data to obtain processed training data;

and carrying out automatic training of the model based on the processed training data to obtain a first inference model.

In an embodiment, when the first determining module 101 is configured to perform automatic training by using the merged training data to obtain the target inference model, it is specifically configured to:

carrying out automatic training by adopting the combined training data to obtain a second reasoning model;

based on the received user-defined feature information and the user-defined weight coefficient, adjusting the combined training data;

and carrying out automatic training by adopting the adjusted training data to obtain a target inference model.

In an embodiment, when the first determining module 101 is configured to perform the adjustment of the merged training data based on the received custom feature information and the custom weight coefficient, specifically, to:

In an embodiment, the second determining module 102 is configured to perform similarity search on the candidate parcel based on preset parcel portrait data to determine a target parcel corresponding to the parcel where the sample object is located in the candidate parcel, and includes:

acquiring basic features constructed based on preset plot portrait data;

extracting candidate features of the candidate land parcels;

and performing feature similarity retrieval on the basic features and the candidate features, and determining a target plot corresponding to the plot where the sample object is located.

In one embodiment, constructing the base features based on the preset plot image data includes:

determining a characteristic value of a preset index based on preset plot image data;

performing down-sampling processing on the characteristic value to obtain a characteristic value after dimensionality reduction;

and analyzing the feature importance based on the feature value after dimension reduction to obtain the basic feature.

In one embodiment, determining the feature value of the predetermined index based on the predetermined parcel image data comprises at least one of:

under a set scale, determining a characteristic value of an adjacent entropy based on the type number of sample objects in an adjacent range of a preset radius of a place where a parcel is located;

In an embodiment, the second determining module 102 is configured to perform feature similarity retrieval on the basic features and the candidate features, and determine a target region corresponding to a region where the sample object is located, including:

performing feature similarity retrieval on the basic features and the candidate features to determine a first land block in the candidate land blocks;

In an embodiment, the first determining module 101, before performing the acquiring of the sample object data related to the selection of the parcel, is further configured to: responding to a plot selection request operation of a client;

the second determining module 102, after performing the determining the target parcel, is further configured to: displaying plot selection information corresponding to a target plot at the client;

the displaying of the parcel selection information corresponding to the target parcel comprises: and displaying the position of the target plot on the map interface based on a preset mark form.

The apparatus according to the embodiment of the present application may execute the method provided by the embodiment of the present application, and the implementation principle is similar, the actions executed by the modules in the apparatus according to the embodiments of the present application correspond to the steps in the method according to the embodiments of the present application, and for the detailed functional description of the modules in the apparatus, reference may be specifically made to the description in the corresponding method shown in the foregoing, and details are not repeated here.

An embodiment of the present application provides an electronic device, including: a memory and a processor; at least one program stored in the memory for execution by the processor, which when executed by the processor, implements: in the application, when sample object data related to plot selection is acquired, firstly, a candidate plot corresponding to a plot where a sample object is located is determined based on preset plot image data and sample object data, at this time, the number of the obtained candidate plots is large, and a plot selection result is relatively rough; on the basis, similarity retrieval is carried out on the candidate plots based on preset plot image data, the plot selection result is further reduced, and the accuracy of plot selection is improved, so that target plots corresponding to the plots where the sample objects are located are determined in the candidate plots; the image data of the land parcel comprises image data corresponding to the land parcel obtained by dividing based on the geographic space. The method and the device for selecting the plot can effectively mine the existing plot portrait data and the sample object data provided by the user, are favorable for improving the reasonability of the plot selection, and enable the selected plot to be more suitable for the user requirements.

In an alternative embodiment, an electronic device is provided, as shown in FIG. 11, the electronic device 1100 shown in FIG. 11 comprising: a processor 1101 and a memory 1103. The processor 1101 is coupled to the memory 1103, such as by a bus 1102. Optionally, the electronic device 1100 may further include a transceiver 1104, and the transceiver 1104 may be used for data interaction between the electronic device and other electronic devices, such as transmission of data and/or reception of data. It should be noted that the transceiver 1104 is not limited to one in practical applications, and the structure of the electronic device 1100 is not limited to the embodiment of the present application.

The Processor 1101 may be a CPU (Central Processing Unit), a general purpose Processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other Programmable logic device, a transistor logic device, a hardware component, or any combination thereof. Which may implement or perform the various illustrative logical blocks, modules, and circuits described in connection with the disclosure. The processor 1101 may also be a combination of computing functions, e.g., comprising one or more microprocessors, DSPs and microprocessors, and the like.

Bus 1102 may include a path that transfers information between the above components. The bus 1102 may be a PCI (Peripheral Component Interconnect) bus, an EISA (Extended Industry Standard Architecture) bus, or the like. The bus 1102 may be divided into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one thick line is shown in FIG. 11, but this is not intended to represent only one bus or type of bus.

The Memory 1103 may be a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, a RAM (Random Access Memory) or other type of dynamic storage device that can store information and instructions, an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical Disc storage, optical Disc storage (including Compact Disc, laser Disc, optical Disc, digital versatile Disc, blu-ray Disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to these.

The memory 1103 is used for storing application program codes (computer programs) for executing the present application, and the execution of the application is controlled by the processor 1101. The processor 1101 is configured to execute application program code stored in the memory 1103 to implement the content shown in the foregoing method embodiments.

Among them, electronic devices include but are not limited to: smart phones, tablet computers, notebook computers, smart speakers, smart watches, vehicle-mounted devices, and the like.

According to an aspect of the application, a computer program product or computer program is provided, comprising computer instructions, the computer instructions being stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the parcel selection method provided in the various alternative implementations described above.

The present application provides a computer-readable storage medium, on which a computer program is stored, which, when running on a computer, enables the computer to execute the corresponding content in the foregoing method embodiments.

It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and may be performed in other orders unless explicitly stated herein. Moreover, at least a portion of the steps in the flow chart of the figure may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

The foregoing is only a partial embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.

27页详细技术资料下载

上一篇：一种医用注射器针头装配设备

下一篇：一种电子档案四性检测实现的方法

Land parcel selection method and device, electronic equipment and storage medium

相关技术

网友询问留言