CN110889029B

CN110889029B - Urban target recommendation method and device

Info

Publication number: CN110889029B
Application number: CN201810946801.5A
Authority: CN
Inventors: 朱翔宇; 翟磊
Original assignee: Jingdong Technology Holding Co Ltd
Current assignee: Jingdong Technology Holding Co Ltd
Priority date: 2018-08-17
Filing date: 2018-08-17
Publication date: 2024-04-05
Anticipated expiration: 2038-08-17
Also published as: CN110889029A

Abstract

The disclosure provides a city target recommendation method, comprising: obtaining city feature data of each type as source data, obtaining a target type to be recommended, taking city feature data which accords with the target type and accords with preset recommendation conditions in the source data as a candidate set, taking city feature data which accords with the target type and does not accord with preset recommendation conditions in the source data as a sample set, training an initial model based on the sample set to obtain an optimal model, obtaining recommendation scores corresponding to the city feature data in the candidate set based on the optimal model, and outputting recommendation results based on the recommendation scores. The disclosure also provides an urban target recommendation device, a computer device and a computer readable medium.

Description

Urban target recommendation method and device

Technical Field

The disclosure relates to the technical field of internet, and more particularly, to a city target recommendation method and device.

Background

The city is used as a background, various resources, information and characteristics in the city are acquired, analyzed and integrated, and corresponding services are provided, so that the city target recommendation system is a basic framework. The existing urban target recommendation schemes are mostly used for manually analyzing various data information, a good automation system is not formed, and the analysis effect is uncontrollable. In addition, since the existing urban target recommendation scheme is often considered only from the dimension of the position information, the decision process is single and incomplete, and the recommendation method used at the same time is usually based on ranking after similarity calculation, wherein a large number of experience parameter modifications are involved, and the dependence on human analysis is high.

Disclosure of Invention

In view of this, the present disclosure provides a more intelligent, rapid and accurate urban target recommendation method and apparatus.

One aspect of the present disclosure provides a city goal recommendation method, comprising: obtaining city feature data of each type as source data, obtaining a target type to be recommended, taking city feature data which accords with the target type and accords with preset recommendation conditions in the source data as a candidate set, taking city feature data which accords with the target type and does not accord with preset recommendation conditions in the source data as a sample set, training an initial model based on the sample set to obtain an optimal model, obtaining recommendation scores corresponding to the city feature data in the candidate set based on the optimal model, and outputting recommendation results based on the recommendation scores.

According to an embodiment of the present disclosure, the acquiring city feature data of each type as the source data includes: dividing the region where the city is located into a plurality of region blocks according to a first preset rule, and dividing the acquired city characteristic data into each region block. In the source data, different types of city feature data within the same region block are associated, and/or the same type of city feature data in different region blocks are associated.

According to an embodiment of the present disclosure, training the initial model based on the sample set to obtain an optimal model includes: adding an output value mark corresponding to the city feature data into each city feature data in the sample set, extracting a plurality of training sets from the sample set according to a second preset rule, wherein each training set comprises a plurality of city feature data with the output value mark, and each training set is different; for each training set, taking part or all of city characteristic data except the training set in the sample set as a verification set corresponding to the training set, training an initial model by using the training set to obtain a training model, verifying the training model by using the verification set, and obtaining an evaluation score of the training model according to a verification result; and taking the training model with the highest evaluation score as an optimal model, and taking a training set corresponding to the training model as an optimal feature combination.

According to an embodiment of the present disclosure, extracting the plurality of training sets from the sample set according to the second preset rule includes: and selecting the city characteristic data for multiple times in a forward and/or backward traversing way by adopting a breadth-first search algorithm in the sample set to obtain each training set.

According to an embodiment of the present disclosure, extracting the plurality of training sets from the sample set according to the second preset rule includes: and for the extracted training set, utilizing a simulated annealing algorithm and/or a random search method to change the training set, and obtaining a new training set.

According to an embodiment of the present disclosure, the obtaining, based on the optimal model, a recommendation score corresponding to each city feature data in the candidate set includes: and inputting the city feature data in the candidate set to the optimal model to obtain corresponding output values, wherein the output value corresponding to each city feature data is a recommendation score corresponding to the city feature data. The outputting the recommendation result based on the recommendation score includes: and sequencing the city feature data from high to low according to the recommendation scores of the city feature data in the candidate set, selecting the city feature data with the preset quantity as recommendation results and outputting the recommendation results.

According to an embodiment of the present disclosure, inputting the city feature data in the candidate set to the optimal model, and obtaining the corresponding output value includes: and inputting the city feature data into the optimal model for multiple times for each city feature data in the candidate set to obtain multiple output values, performing cross-validation optimization on the multiple output values to obtain an optimized output value, and taking the optimized output value as a recommendation score of the city feature data.

According to an embodiment of the present disclosure, the acquiring city feature data of each type as the source data includes: and cleaning the acquired city characteristic data, wherein the city characteristic data after data cleaning form source data.

Another aspect of the present disclosure provides an urban target recommendation device, including: the system comprises a first acquisition module, a second acquisition module and a recommendation processing module. The first acquisition module is used for acquiring city characteristic data of each type as source data. The second acquisition module is used for acquiring the type of the target to be recommended. The recommendation processing module is used for taking city feature data which accords with the target type and accords with preset recommendation conditions in source data as a candidate set, taking city feature data which accords with the target type and does not accord with preset recommendation conditions in the source data as a sample set, training an initial model based on the sample set to obtain an optimal model, obtaining recommendation scores corresponding to the city feature data in the candidate set based on the optimal model, and outputting recommendation results based on the recommendation scores.

According to an embodiment of the present disclosure, the first obtaining module obtaining city feature data of each type as source data includes: the first acquisition module is used for dividing the area where the city is located into a plurality of area blocks according to a first preset rule; dividing the acquired city characteristic data into area blocks; in the source data, different types of city feature data within the same region block are associated, and/or the same type of city feature data in different region blocks are associated.

According to an embodiment of the disclosure, the recommendation processing module trains the initial model based on the sample set, and the obtaining of the optimal model includes: the recommendation processing module is used for adding an output value mark corresponding to the city characteristic data in each city characteristic data in the sample set; extracting a plurality of training sets from the sample set according to a second preset rule, wherein each training set comprises a plurality of pieces of city characteristic data with output value marks, and the training sets are different; for each training set, taking part or all of city characteristic data except the training set in the sample set as a verification set corresponding to the training set, training an initial model by using the training set to obtain a training model, verifying the training model by using the verification set, and obtaining an evaluation score of the training model according to a verification result; and taking the training model with the highest evaluation score as an optimal model, and taking a training set corresponding to the training model as an optimal feature combination.

According to an embodiment of the present disclosure, the recommendation processing module extracting a plurality of training sets from the sample set according to a second preset rule comprises: the recommendation processing module is used for selecting the city feature data for multiple times in a forward and/or backward traversal mode by adopting a breadth-first search algorithm in the sample set to obtain each training set.

According to an embodiment of the present disclosure, the recommendation processing module extracting a plurality of training sets from the sample set according to a second preset rule comprises: and the recommendation processing module is used for changing the extracted training set by using a simulated annealing algorithm and/or a random search method to obtain a new training set.

According to an embodiment of the present disclosure, the recommendation processing module obtaining recommendation scores corresponding to the city feature data in the candidate set based on the optimal model includes: and the recommendation processing module is used for inputting the urban feature data in the candidate set into the optimal model to obtain corresponding output values, and the output value corresponding to each urban feature data is a recommendation score corresponding to the urban feature data. The recommendation processing module outputting a recommendation result based on the recommendation score comprises: and the recommendation processing module is used for sequencing the urban characteristic data from high to low according to the recommendation scores of the urban characteristic data in the candidate set, selecting the urban characteristic data with the preset quantity before as recommendation results and outputting the recommendation results.

According to an embodiment of the disclosure, the recommendation processing module inputs each city feature data in the candidate set to the optimal model, and obtaining the corresponding output value includes: the recommendation processing module is used for inputting the city characteristic data into the optimal model for a plurality of times for each city characteristic data in the candidate set to obtain a plurality of output values; and performing cross-validation optimization on the plurality of output values to obtain optimized output values, and taking the optimized output values as recommendation scores of the city feature data.

According to an embodiment of the present disclosure, the first obtaining module obtaining city feature data of each type as source data includes: the first acquisition module is used for carrying out data cleaning on the acquired city characteristic data, and the city characteristic data after data cleaning form source data.

Another aspect of the present disclosure provides a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the method as described above when executing the program.

Another aspect of the present disclosure provides a computer readable medium having stored thereon executable instructions which, when executed by a processor, cause the processor to perform a method as described above.

Another aspect of the present disclosure provides a non-volatile storage medium storing computer executable instructions that when executed are configured to implement a method as described above.

Another aspect of the present disclosure provides a computer program comprising computer executable instructions which when executed are for implementing a method as described above.

According to the embodiment of the disclosure, the problems that a good automation system is not formed, the effect is uncontrollable, the decision process is single and incomplete, and the dependence on human analysis is high in the existing urban target recommendation field can be at least partially solved/lightened/inhibited/even avoided, model training is performed on the basis of a sample set in source data, data to be recommended in a candidate set in the source data is automatically evaluated by utilizing an optimal model obtained through training, and then the urban target recommendation result can be output according to the evaluation result, wherein the optimal model obtained through sample set training is adapted to the candidate set because the sample set and the candidate set correspond to the same urban target type, recommendation scores of urban characteristic data in the candidate set can be intelligently, rapidly and accurately evaluated by utilizing the optimal model, and accordingly, more accurate and reliable recommendation results can be obtained.

Drawings

The above and other objects, features and advantages of the present disclosure will become more apparent from the following description of embodiments thereof with reference to the accompanying drawings in which:

FIG. 1 schematically illustrates an exemplary system architecture to which city goal recommendation methods and apparatus may be applied, in accordance with embodiments of the present disclosure;

FIG. 2 schematically illustrates a flow chart of a city goal recommendation method, according to an embodiment of the present disclosure;

FIG. 3 schematically illustrates a schematic diagram of constructing source data according to an embodiment of the present disclosure;

FIG. 4 schematically illustrates a flow chart of a method of model training in accordance with an embodiment of the present disclosure;

FIG. 5 schematically illustrates a schematic diagram of training set iterative optimization in accordance with an embodiment of the present disclosure;

FIG. 6 schematically illustrates a block diagram of a city goal recommendation device, in accordance with an embodiment of the present disclosure; and

fig. 7 schematically illustrates a block diagram of a computer device according to an embodiment of the disclosure.

Detailed Description

Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. It should be understood that the description is only exemplary and is not intended to limit the scope of the present disclosure. In the following detailed description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the present disclosure. It may be evident, however, that one or more embodiments may be practiced without these specific details. In addition, in the following description, descriptions of well-known structures and techniques are omitted so as not to unnecessarily obscure the concepts of the present disclosure.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The terms "comprises," "comprising," and/or the like, as used herein, specify the presence of stated features, steps, operations, and/or components, but do not preclude the presence or addition of one or more other features, steps, operations, or components.

All terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. It should be noted that the terms used herein should be construed to have meanings consistent with the context of the present specification and should not be construed in an idealized or overly formal manner.

Where expressions like at least one of "A, B and C, etc. are used, the expressions should generally be interpreted in accordance with the meaning as commonly understood by those skilled in the art (e.g.," a system having at least one of A, B and C "shall include, but not be limited to, a system having a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.). Where a formulation similar to at least one of "A, B or C, etc." is used, in general such a formulation should be interpreted in accordance with the ordinary understanding of one skilled in the art (e.g. "a system with at least one of A, B or C" would include but not be limited to systems with a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.). It should also be appreciated by those skilled in the art that virtually any disjunctive word and/or phrase presenting two or more alternative items, whether in the description, claims, or drawings, should be understood to contemplate the possibilities of including one of the items, either of the items, or both. For example, the phrase "a or B" should be understood to include the possibility of "a" or "B", or "a and B".

The embodiment of the disclosure provides a city target recommending method and device. The method comprises a source data acquisition process, a target type acquisition process to be recommended, a model training process, a model prediction process and a recommendation result output process. The method comprises the steps of extracting a sample set and a candidate set from source data based on a target type to be recommended, performing model training process based on the sample set, performing predictive scoring on data in the candidate set based on an optimal model obtained through training, and outputting a corresponding recommendation result based on the predictive scoring.

Fig. 1 schematically illustrates an exemplary system architecture 100 to which city goal recommendation methods and apparatus may be applied, in accordance with embodiments of the present disclosure. It should be noted that fig. 1 is only an example of a system architecture to which embodiments of the present disclosure may be applied to assist those skilled in the art in understanding the technical content of the present disclosure, but does not mean that embodiments of the present disclosure may not be used in other devices, systems, environments, or scenarios.

As shown in fig. 1, a system architecture 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 is used as a medium to provide communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, among others.

The user may interact with the server 105 via the network 104 using the terminal devices 101, 102, 103 to receive or send messages or the like. Various client applications, such as shopping class applications, web browser applications, target recommendation class applications, search recommendation class applications, instant messaging tools, mailbox clients, social platform software, etc. (by way of example only) may be installed on the terminal devices 101, 102, 103.

The terminal devices 101, 102, 103 may be a variety of electronic devices having a display screen and supporting web browsing, including but not limited to smartphones, tablets, laptop and desktop computers, and the like.

The server 105 may be a server providing various services, such as a background management server (by way of example only) providing support for websites browsed by users using the terminal devices 101, 102, 103. The background management server may analyze and process the received data such as the urban target recommendation request and the target type to be recommended, and feed back the processing result (such as a recommendation result, a web page, information, or data obtained or generated according to the user request) to the terminal device.

It should be noted that, the urban objective recommendation method provided by the embodiments of the present disclosure may be performed by the server 105. Accordingly, the city goal recommendation device provided by the embodiments of the present disclosure may be generally disposed in the server 105. The city goal recommendation method provided by the embodiments of the present disclosure may also be performed by a server or a cluster of servers that are different from the server 105 and that are capable of communicating with the terminal devices 101, 102, 103 and/or the server 105. Accordingly, the city goal recommendation apparatus provided by the embodiments of the present disclosure may also be provided in a server or a server cluster that is different from the server 105 and is capable of communicating with the terminal devices 101, 102, 103 and/or the server 105.

Alternatively, the urban target recommendation method provided by the embodiments of the present disclosure may also be performed by the terminal devices 101, 102, 103. Accordingly, the urban target recommendation device provided by the embodiments of the present disclosure may be generally provided in the terminal devices 101, 102, 103. The city goal recommendation method provided by the embodiments of the present disclosure may also be performed by a server or a server cluster that is different from the terminal devices 101, 102, 103 and that is capable of communicating with the terminal devices 101, 102, 103 and/or the server 105. Accordingly, the city target recommendation apparatus provided by the embodiments of the present disclosure may also be provided in a server or a server cluster different from the terminal devices 101, 102, 103 and capable of communicating with the terminal devices 101, 102, 103 and/or the server 105.

It should be understood that the number of terminal devices, networks, and servers in fig. 1 is merely illustrative, and that any number of terminal devices, networks, and servers may be provided as desired for implementation.

Fig. 2 schematically illustrates a flowchart of a city goal recommendation method, according to an embodiment of the present disclosure.

As shown in fig. 2, the method includes acquiring city feature data of each type as source data in operation S210.

The city feature data of each type in the present operation may be related feature data of each type of object in a specified city, for example, feature data of shops distributed in the city, feature data of residences distributed in the city, feature data of greenbelts distributed in the city, etc., and without limitation, feature data of any city object that may have recommended needs may be used as source data of the present scheme.

Then, in operation S220, a target type to be recommended is acquired.

In operation S230, the city feature data in the source data that meets the target type and meets the preset recommendation condition is used as a candidate set, and the city feature data in the source data that meets the target type and does not meet the preset recommendation condition is used as a sample set.

In this operation, the preset recommendation condition refers to a condition that whether the city feature data can be listed as a candidate recommendation target is judged under the condition that the types are in line, the preset recommendation condition can be different under different conditions, the preset recommendation condition needs to be determined according to the user requirement corresponding to the target to be recommended, the purpose is to take the city feature data which is in line with the recommendation target type and can be directly used by the user according to the user requirement as a candidate set, the city feature data which is in line with the recommendation target type and cannot be used by the user can be used as a sample set to provide reference for evaluating the city feature data in the candidate set.

In operation S240, the initial model is trained based on the sample set to obtain an optimal model.

In operation S250, recommendation scores corresponding to the city feature data in the candidate set are obtained based on the optimal model.

In operation S260, a recommendation result is output based on the recommendation score.

It should be noted that, the urban objective recommendation method shown in fig. 2 may be implemented at the server side or at the client side, when the method is implemented at the server side, source data is acquired at operation S210, the type of objective to be recommended submitted by the client may be received at operation S220, a sample set and a candidate set are divided from the source data at operations S230 to S250, an optimal model is obtained according to training of the sample set, recommendation scores of urban feature data in the candidate set are obtained according to the optimal model, a recommendation result is output to the client based on the recommendation scores at operation S260, and the recommendation result is output and displayed to the user by the client. When the method is implemented at the client side, source data is acquired in operation S210, a target type to be recommended input or selected by a user can be received in operation S220, a sample set and a candidate set are divided from the source data in operations S230 to S250, an optimal model is obtained according to training of the sample set, recommendation scores of the feature data of each city in the candidate set are obtained according to the optimal model, and a recommendation result is output to the user based on the recommendation scores in operation S260.

Therefore, the method shown in fig. 2 performs model training based on the sample set in the source data, automatically evaluates the data to be recommended in the candidate set in the source data by using the optimal model obtained by training, and further can output the urban target recommendation result according to the evaluation result, wherein the optimal model obtained by training the sample set is adapted to the candidate set because the sample set and the candidate set correspond to the same urban target type, and the recommendation scores of the urban feature data in the candidate set can be evaluated intelligently, rapidly and accurately by using the optimal model, so that more accurate and reliable recommendation results can be obtained.

In one embodiment of the present disclosure, the step S210 of the method shown in fig. 2 of obtaining city feature data of each type as source data includes: dividing the area where the city is located into a plurality of area blocks according to a first preset rule, dividing the acquired city feature data into each area block, and associating different types of city feature data in the same area block and/or associating the same types of city feature data in different area blocks in the source data. The method comprises the steps that a city is a specified target city for target recommendation, the city characteristic data are divided into regional blocks according to position information corresponding to the city characteristic data, source data can be recorded based on the association between different regional blocks to which the city characteristic data of the same type belong, the source data can be recorded based on the association between the city characteristic data of different types in the same regional block, the source data can describe the target city according to the division angle of the regional blocks, the city can be called as a transverse characteristic system, the target city can be described according to the division angle of the types of the city characteristic data, the city can be called as a longitudinal characteristic system, thus the source data can describe the characteristics of the target city more comprehensively, integrally and abundantly from different angles, the data characteristic support is provided for multidimensional recommendation in the smart city, the transverse characteristic system is described by the whole angle of the city, and the longitudinal characteristic system is described by the direction of the richness of regional content.

FIG. 3 schematically illustrates a schematic diagram of constructing source data according to an embodiment of the present disclosure. The target city is a city A, and city target recommendation is required to be performed in the city A, so that city characteristic data corresponding to the city A is required to be obtained as source data. As shown in fig. 3, the area where the city a is located is divided into 9 area blocks, and the longitude and latitude corresponding to each area block can be converted into a character string through a GeoHash algorithm to be used as the codes corresponding to each area block. The acquired city feature data of the city a is divided into each regional block according to the respective position information, and the same type of city feature data in different regional blocks are associated, and in fig. 3, a set of the city feature data of the first type, a set of the city feature data of the second type, … …, and a set of the city feature data of the nth type distributed in 9 regional blocks are shown, and are respectively labeled as a first feature group, a second feature group, … …, and an nth feature group in the figure, and a plurality of feature groups form all source data. In this example, dividing the city into a plurality of regional blocks may integrate city feature information of too dense different points in the city, and aggregate the scattered point features in the city into regional features from a spatial perspective, so as to better describe the target city. Each area may include a feature group composed of city feature data of various types, including, but not limited to, business district feature groups, guest group feature groups, traffic feature groups, time period feature groups, etc., which are not limited herein, for the purpose of enriching city feature information of various areas and constructing a city feature data network.

Further, in one embodiment of the present disclosure, to ensure validity of the city feature data in the source data, the obtaining, by operation S210 of the method shown in fig. 2, city feature data of each type as the source data includes: and cleaning the acquired city characteristic data, removing dirty data, expired data, redundant data and the like, and forming source data based on the city characteristic data after data cleaning.

Fig. 4 schematically illustrates a flowchart of a method of model training according to an embodiment of the present disclosure, to illustrate a process of training an initial model based on the sample set to obtain an optimal model in operation S240 in the method illustrated in fig. 2, and the implementation of the method illustrated in fig. 4 is after the division of the candidate set and the sample set in the source data is completed in operation S230 in the method illustrated in fig. 2.

As shown in fig. 4, the method includes adding an output value tag corresponding to each city feature data in the sample set in operation S241.

The operation adds output value marks for each city characteristic data in the sample set, so that the subsequent model training process is a supervised learning process.

In operation S242, a plurality of training sets, each including a plurality of pieces of city feature data with output value markers, are extracted from the sample set according to a second preset rule, and the training sets are different.

In operation S243, for each training set, part or all of city feature data in the sample set except the training set is used as a verification set corresponding to the training set, the training set is used to train the initial model to obtain a training model, the verification set is used to verify the training model, and an evaluation score of the training model is obtained according to the verification result.

In operation S244, the training model with the highest evaluation score is used as the optimal model, and the training set corresponding to the training model is used as the optimal feature combination.

Then, based on the method shown in fig. 4, the trained optimal model may predict the output value of the city feature data in the candidate set, and in one embodiment of the present disclosure, operation S250 of the method shown in fig. 2, obtaining, based on the optimal model, a recommendation score corresponding to each city feature data in the candidate set includes: and inputting the city feature data in the candidate set to the optimal model to obtain corresponding output values, wherein the output value corresponding to each city feature data is a recommendation score corresponding to the city feature data. Operation S260 outputting the recommendation result based on the recommendation score includes: and sequencing the city feature data from high to low according to the recommendation scores of the city feature data in the candidate set, selecting the city feature data with the preset quantity as recommendation results and outputting the recommendation results.

In order to further optimize the prediction result of the optimal model on the urban feature data in the candidate set, as an optional embodiment, the inputting each urban feature data in the candidate set into the optimal model, and obtaining the corresponding output value includes: and inputting the city characteristic data into the optimal model for multiple times for each city characteristic data in the candidate set to obtain multiple output values. And performing cross-validation optimization on the plurality of output values to obtain optimized output values, and taking the optimized output values as recommendation scores of the city feature data. In this embodiment, after the optimal model is found, in order to avoid instability of a single output structure, stability optimization is performed on the output result, so that recommendation scores corresponding to the feature data of each city in the candidate set are more accurate and reliable.

It can be seen that the method shown in fig. 4 is expanded to illustrate operation S240 shown in fig. 2, different training models are obtained by continuously transforming combinations of training sets in a sample set, and then the advantages and disadvantages of the obtained training models are verified by a verification set in the sample set, so that an optimal model and a corresponding optimal feature combination can be effectively found, a subsequent recommendation result is more accurate, and a complete process of urban target recommendation can be realized by combining other operation processes in fig. 2. The following is a specific example:

The method comprises the steps of carrying out city target recommendation by taking a city A as a target city, firstly acquiring various types of city characteristic data about the city A, cleaning the acquired city characteristic data, taking the cleaned city characteristic data as source data, firstly extracting a longitudinal characteristic system in each regional block divided in the city A, and then constructing a transverse characteristic system among regional blocks on each type of city characteristic data, so that the source data can link all information points in the city A into pieces in the form of regional blocks, and aggregating different regional blocks into the characteristic system for describing the characteristics of the whole city by constructing the transverse characteristic system. And then acquiring the type of the target to be recommended, wherein the type of the target to be recommended defines the recommendation problem to be solved in the range of the city A, including but not limited to shop site selection recommendation, operation scheme recommendation and the like, and the purpose is to provide a necessary scheme for the subsequent design of a basic model for solving the recommendation problem through good definition of the recommendation problem. Taking a shop site selection problem as an example, the type of the target to be recommended in the example is a shop, and the corresponding user requirement can be that the shop which is most suitable for business is found through the urban target recommendation of the shop type. According to the type of the target to be recommended, corresponding candidate sets and sample sets are extracted from the source data, in the case of the example, when the type of the target to be recommended is a shop, the preset recommendation condition can be in a renting state, city feature data which belongs to the shop type and is in the renting state in the source data is taken as the candidate set, and city feature data which belongs to the shop type but is not in the renting state in the source data is taken as the sample set. Constructing a basic model for solving a specific recommendation problem corresponding to a target to be recommended, specifically constructing a basic model for solving the recommended target from the aspects of characteristics and algorithms, and taking the basic model as an initial model for subsequent training.

In this example, each city feature data in the sample set is city feature data belonging to a shop type but not in a renting state, for example, the city feature data may be related to a shop in a business state, each city feature data in the sample set may include related information such as location information, passenger flow information, traffic condition information, surrounding environment information and the like of the corresponding shop, an output value mark corresponding to the city feature information may be calculated according to the related information, and the output value mark may be used as an index for evaluating the business condition of the corresponding shop.

And extracting different training sets from the sample set added with the output value marks for a plurality of times to train the initial model so as to obtain an optimal model. The training sets can be selected for multiple times in a sample set in a forward and/or backward searching mode, specifically, the width-first searching algorithm is adopted in the sample set to conduct forward and/or backward traversing on city feature data, so that each training set is obtained, for each training set, part or all of city feature data except the training set in the sample set is used as a verification set corresponding to the training set, an initial model is trained by using the training set to obtain a training model, the verification set is used for verifying the training model, the evaluation score of the training model is obtained according to a verification result, the training model with the highest evaluation score is used as an optimal model, and the training set corresponding to the training model is used as an optimal feature combination. The process can be performed in a manner that, for a training set T1 extracted for the first time, part or all of city feature data in the sample set except the extracted training set T1 is used as a verification set V1 corresponding to the training set Tl, the training set T1 is used for training the initial model to obtain a training model M1, the corresponding verification set V1 is used for verifying the training model M1 obtained through training, an evaluation score of the training model M1 is obtained according to the verification result, whether the training model M1 is optimized relative to the initial model is judged through the evaluation score, if so, the training set T2 is obtained through continuing to extract the training set on the basis of the training set T1, part or all of city feature data in the sample set except the training set T2 is used as a verification set V2, the training set T2 is used for training on the basis of the initial model or the training model M1 to obtain a training model M2, the corresponding verification set V2 is used for verifying the training model M2 obtained through the corresponding verification set, the evaluation score of the training model M2 is obtained according to the verification result, if the training model M2 is judged to be optimized relative to the initial model, if so, if the optimal score is found on the basis, and the optimal score is found, and the training set is continued on the basis of the evaluation. In the above process, if the model corresponding to the newly extracted training set is inferior to the model corresponding to the previously extracted training set, the search scan needs to be performed again from the sample set to extract the new training set.

Fig. 5 schematically illustrates a schematic diagram of training set iterative optimization in accordance with an embodiment of the present disclosure. In the above process, the training set is continuously and iteratively optimized to finally find the optimal feature combination, and the optimal model can be obtained by training using the optimal feature combination, as can be seen from fig. 5, the training set extracted from the sample set of the source data gradually becomes smaller in the iterative optimization process, so as to form a tower-shaped structure, which can be called a city data tower, and represents the optimization process of sample data required for solving the problem of the corresponding target to be recommended in the city a. Any city goal recommendation problem can ultimately be solved by constructing such a corresponding city data tower.

Further, as an optional embodiment, when the effect of the training model obtained by the training is not improved, the current training set may be changed by using a simulated annealing algorithm and a random search method to obtain a new training set, and whether the new training set is optimized with respect to the previous training set is judged, and if so, the process of optimizing the training set may be repeated on the basis of the new training set until the training set obtained by adjustment is not further optimized by the training model. By adding the simulated annealing algorithm and the random search method, the possibility of finding the optimal combination in the iterative optimization process of the training set is improved, namely when the iterative optimization process cannot continue to obtain the training model with better effect, one or more city feature data in the training set are randomly replaced or one or more city feature combinations are randomly added in the training set, so that the problem that the local optimal feature combination possibly falls into in the iterative optimization process of the training set can be avoided as much as possible.

And after the optimal model corresponding to the optimal feature combination is found, greedy cross-validation can be performed on the optimal model, specifically, the sample set is cut to obtain a plurality of subsets, and the optimal model is cross-validated by the plurality of subsets, so that the occurrence of contingency in model training is avoided, and the optimal model with better stability is trained.

In this example, after the optimal model corresponding to the optimal feature model is obtained, each city feature data in the candidate set is input to the optimal model, the optimal model outputs an output value corresponding to each city feature data, the output value is used as a recommendation score of the corresponding city feature data, the city feature data in the candidate set is sequenced from high to low according to the recommendation score, the preset number of city feature data is used as a recommendation result to be output, and the shop corresponding to the preset number of city recommendation data is recommended to the user for the user to perform shop site selection. Other types of city target recommendations are the same and are not described in detail.

Fig. 6 schematically illustrates a block diagram of a city goal recommendation device, in accordance with an embodiment of the present disclosure.

As shown in fig. 6, the article searching apparatus 600 includes a first acquisition module 610, a second acquisition module 620, and a recommendation processing module 630.

The first obtaining module 610 is configured to obtain city feature data of each type as source data.

The second obtaining module 620 is configured to obtain a type of an object to be recommended.

The recommendation processing module 630 is configured to take, as a candidate set, city feature data in the source data, which meets the target type and meets a preset recommendation condition, take, as a sample set, city feature data in the source data, which meets the target type and does not meet the preset recommendation condition, train an initial model based on the sample set, obtain an optimal model, obtain recommendation scores corresponding to the city feature data in the candidate set based on the optimal model, and output a recommendation result based on the recommendation scores.

Therefore, the device shown in fig. 6 performs model training based on the sample set in the source data, and automatically evaluates the data to be recommended in the candidate set in the source data by using the optimal model obtained by training, so that the urban target recommendation result can be output according to the evaluation result, wherein the optimal model obtained by training the sample set is adapted to the candidate set because the sample set and the candidate set correspond to the same urban target type, and recommendation scores of the urban feature data in the candidate set can be intelligently, rapidly and accurately evaluated by using the optimal model, so that more accurate and reliable recommendation results can be obtained.

In one embodiment of the present disclosure, the first obtaining module 610 obtains each type of city feature data as source data includes: the first obtaining module 610 is configured to divide the area where the city is located into a plurality of area blocks according to a first preset rule, and divide the obtained city feature data into the area blocks. In the source data, different types of city feature data within the same region block are associated, and/or the same type of city feature data in different region blocks are associated.

In one embodiment of the present disclosure, the recommendation processing module 630 trains the initial model based on the sample set, and the obtaining the optimal model includes: the recommendation processing module 630 is configured to add an output value tag corresponding to each city feature data in the sample set, extract a plurality of training sets from the sample set according to a second preset rule, where each training set includes a plurality of city feature data with output value tags, and each training set is different; and for each training set, taking part or all of city characteristic data except the training set in the sample set as a verification set corresponding to the training set, training an initial model by using the training set to obtain a training model, verifying the training model by using the verification set, obtaining an evaluation score of the training model according to a verification result, taking the training model with the highest evaluation score as an optimal model, and taking the training set corresponding to the training model as an optimal characteristic combination.

Wherein, as an optional embodiment, the recommendation processing module 630 extracts a plurality of training sets from the sample set according to a second preset rule includes: the recommendation processing module 630 is configured to perform multiple selections on the city feature data in the sample set by using a breadth-first search algorithm to traverse forward and/or backward, so as to obtain each training set.

As another alternative embodiment, recommendation processing module 630 extracts a plurality of training sets from the sample set according to a second preset rule comprises: the recommendation processing module 630 is configured to change the extracted training set by using a simulated annealing algorithm and/or a random search method, so as to obtain a new training set.

In one embodiment of the present disclosure, the recommendation processing module 630 obtains recommendation scores corresponding to each city feature data in the candidate set based on the optimal model includes: the recommendation processing module 630 is configured to input each city feature data in the candidate set to the optimal model, and obtain a corresponding output value, where the output value corresponding to each city feature data is a recommendation score corresponding to the city feature data. And, the recommendation processing module 630 outputting recommendation results based on the recommendation scores includes: the recommendation processing module 630 is configured to sort the city feature data according to the recommendation scores of the city feature data in the candidate set from high to low, select a preset number of city feature data as recommendation results, and output the selected city feature data.

Specifically, the recommendation processing module 630 inputs each city feature data in the candidate set to the optimal model, and obtaining the corresponding output value may include: the recommendation processing module 630 is configured to input, for each city feature data in the candidate set, the city feature data into the optimal model multiple times, obtain multiple output values, perform cross-validation optimization on the multiple output values, obtain an optimized output value, and use the optimized output value as a recommendation score of the city feature data.

As an alternative embodiment, the first obtaining module 610 obtains city feature data of each type as source data includes: the first obtaining module 610 is configured to perform data cleaning on the obtained city feature data, where the city feature data after data cleaning constitutes source data.

It should be noted that, the city target recommending apparatus 600 shown in fig. 6 may be configured on the server side or on the client side, when the city target recommending apparatus 600 is configured on the server side, the first acquiring module 610 acquires source data, the second acquiring module 620 may receive a target type to be recommended submitted by the client, the recommendation processing module 630 divides a sample set and a candidate set from the source data, trains according to the sample set to obtain an optimal model, obtains recommendation scores of city feature data in the candidate set according to the optimal model, outputs a recommendation result to the client based on the recommendation scores, and outputs and displays the recommendation result to the user by the client. When the urban target recommendation device 600 is configured on the client side, the first acquiring module 610 acquires the source data, the second acquiring module 620 may receive the target type to be recommended input or selected by the user, the recommendation processing module 630 divides the source data into a sample set and a candidate set, trains the sample set to obtain an optimal model, obtains recommendation scores of the urban feature data in the candidate set according to the optimal model, and outputs a recommendation result to the user based on the recommendation scores in operation S260.

It should be noted that, in the embodiment of the apparatus portion, the implementation manner, the solved technical problem, the implemented function, and the achieved technical effect of each module/unit/subunit and the like are the same as or similar to the implementation manner, the solved technical problem, the implemented function, and the achieved technical effect of each corresponding step in the embodiment of the method portion, and are not described herein again.

Any number of modules, sub-modules, units, sub-units, or at least some of the functionality of any number of the sub-units according to embodiments of the present disclosure may be implemented in one module. Any one or more of the modules, sub-modules, units, sub-units according to embodiments of the present disclosure may be implemented as split into multiple modules. Any one or more of the modules, sub-modules, units, sub-units according to embodiments of the present disclosure may be implemented at least in part as a hardware circuit, such as a Field Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a system-on-chip, a system-on-substrate, a system-on-package, an Application Specific Integrated Circuit (ASIC), or in any other reasonable manner of hardware or firmware that integrates or encapsulates the circuit, or in any one of or a suitable combination of three of software, hardware, and firmware. Alternatively, one or more of the modules, sub-modules, units, sub-units according to embodiments of the present disclosure may be at least partially implemented as computer program modules, which when executed, may perform the corresponding functions.

For example, any of the first acquisition module 610, the second acquisition module 620, and the recommendation processing module 630 may be combined in one module to be implemented, or any of the modules may be split into a plurality of modules. Alternatively, at least some of the functionality of one or more of the modules may be combined with at least some of the functionality of other modules and implemented in one module. According to embodiments of the present disclosure, at least one of the first acquisition module 610, the second acquisition module 620, and the recommendation processing module 630 may be implemented at least in part as hardware circuitry, such as a Field Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a system on a chip, a system on a substrate, a system on a package, an Application Specific Integrated Circuit (ASIC), or may be implemented in hardware or firmware in any other reasonable way of integrating or packaging circuitry, or in any one of or a suitable combination of three of software, hardware, and firmware. Alternatively, at least one of the first acquisition module 610, the second acquisition module 620, and the recommendation processing module 630 may be at least partially implemented as a computer program module, which when executed, may perform the corresponding functions.

Fig. 7 schematically illustrates a block diagram of a computer device adapted to implement the above-described method according to an embodiment of the present disclosure. The computer device illustrated in fig. 7 is merely an example and should not be construed as limiting the functionality and scope of use of embodiments of the present disclosure.

As shown in fig. 7, a computer device 700 according to an embodiment of the present disclosure includes a processor 701 that can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) 702 or a program loaded from a storage section 708 into a Random Access Memory (RAM) 703. The processor 701 may include, for example, a general purpose microprocessor (e.g., a CPU), an instruction set processor and/or an associated chipset and/or a special purpose microprocessor (e.g., an Application Specific Integrated Circuit (ASIC)), or the like. The processor 701 may also include on-board memory for caching purposes. The processor 701 may comprise a single processing unit or a plurality of processing units for performing different actions of the method flows according to embodiments of the disclosure.

In the RAM 703, various programs and data required for the operation of the computer device 700 are stored. The processor 701, the ROM 702, and the RAM 703 are connected to each other through a bus 704. The processor 701 performs various operations of the method flow according to the embodiments of the present disclosure by executing programs in the ROM 702 and/or the RAM 703. Note that the program may be stored in one or more memories other than the ROM 702 and the RAM 703. The processor 701 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in the one or more memories.

According to an embodiment of the present disclosure, the computer device 700 may also include an input/output (I/O) interface 705, the input/output (I/O) interface 705 also being connected to the bus 704. The system 700 may also include one or more of the following components connected to the I/O interface 705: an input section 706 including a keyboard, a mouse, and the like; an output portion 707 including a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, a speaker, and the like; a storage section 708 including a hard disk or the like; and a communication section 709 including a network interface card such as a LAN card, a modem, or the like. The communication section 709 performs communication processing via a network such as the internet. The drive 710 is also connected to the I/O interface 705 as needed. A removable medium 711 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 710 as necessary, so that a computer program read therefrom is mounted into the storage section 708 as necessary.

According to embodiments of the present disclosure, the method flow according to embodiments of the present disclosure may be implemented as a computer software program. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method shown in the flowcharts. In such an embodiment, the computer program may be downloaded and installed from a network via the communication portion 709, and/or installed from the removable medium 711. The above-described functions defined in the system of the embodiments of the present disclosure are performed when the computer program is executed by the processor 701. The systems, devices, apparatus, modules, units, etc. described above may be implemented by computer program modules according to embodiments of the disclosure.

The present disclosure also provides a computer-readable medium that may be embodied in the apparatus/device/system described in the above embodiments; or may exist alone without being assembled into the apparatus/device/system. The computer readable medium carries one or more programs which, when executed, implement methods in accordance with embodiments of the present disclosure.

According to embodiments of the present disclosure, the computer readable medium may be a computer readable signal medium or a computer readable storage medium or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this disclosure, a computer-readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present disclosure, however, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, fiber optic cable, radio frequency signals, or the like, or any suitable combination of the foregoing.

For example, according to embodiments of the present disclosure, the computer-readable medium may include ROM 702 and/or RAM 703 and/or one or more memories other than ROM 702 and RAM 703 described above.

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

Those skilled in the art will appreciate that the features recited in the various embodiments of the disclosure and/or in the claims may be provided in a variety of combinations and/or combinations, even if such combinations or combinations are not explicitly recited in the disclosure. In particular, the features recited in the various embodiments of the present disclosure and/or the claims may be variously combined and/or combined without departing from the spirit and teachings of the present disclosure. All such combinations and/or combinations fall within the scope of the present disclosure.

The embodiments of the present disclosure are described above. However, these examples are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although the embodiments are described above separately, this does not mean that the measures in the embodiments cannot be used advantageously in combination. The scope of the disclosure is defined by the appended claims and equivalents thereof. Various alternatives and modifications can be made by those skilled in the art without departing from the scope of the disclosure, and such alternatives and modifications are intended to fall within the scope of the disclosure.

Claims

1. A city goal recommendation method, comprising:

obtaining city characteristic data of each type as source data;

obtaining a target type to be recommended;

Taking city characteristic data which accords with the target type and accords with preset recommendation conditions in the source data as a candidate set, and taking city characteristic data which accords with the target type and does not accord with the preset recommendation conditions in the source data as a sample set;

adding an output value mark corresponding to the city feature data into each city feature data in the sample set to obtain the sample set added with the output value mark;

training an initial model based on the sample set added with the output value mark to obtain an optimal model;

obtaining recommendation scores corresponding to the city feature data in the candidate set based on the optimal model;

and outputting a recommendation result based on the recommendation score.

2. The method of claim 1, wherein the obtaining city feature data of each type as source data comprises:

dividing the area where the city is located into a plurality of area blocks according to a first preset rule;

dividing the acquired city characteristic data into area blocks;

in the source data, different types of city feature data within the same region block are associated, and/or the same type of city feature data in different region blocks are associated.

3. The method of claim 1, wherein the training an initial model based on the sample set to which the output value markers have been added to obtain an optimal model comprises:

extracting a plurality of training sets from the sample set according to a second preset rule, wherein each training set comprises a plurality of pieces of city characteristic data with output value marks, and the training sets are different;

for each training set, taking part or all of city characteristic data except the training set in the sample set as a verification set corresponding to the training set, training an initial model by using the training set to obtain a training model, verifying the training model by using the verification set, and obtaining an evaluation score of the training model according to a verification result;

and taking the training model with the highest evaluation score as an optimal model, and taking a training set corresponding to the training model as an optimal feature combination.

4. A method according to claim 3, wherein the extracting a plurality of training sets from the sample set according to a second preset rule comprises:

and selecting the city characteristic data for multiple times in a forward and/or backward traversing way by adopting a breadth-first search algorithm in the sample set to obtain each training set.

5. A method according to claim 3, wherein the extracting a plurality of training sets from the sample set according to a second preset rule comprises:

and for the extracted training set, utilizing a simulated annealing algorithm and/or a random search method to change the training set, and obtaining a new training set.

6. A method according to claim 3, wherein:

the obtaining the recommendation scores corresponding to the city feature data in the candidate set based on the optimal model comprises the following steps: inputting each city characteristic data in the candidate set into the optimal model to obtain corresponding output values, wherein the output value corresponding to each city characteristic data is a recommendation score corresponding to the city characteristic data;

the outputting the recommendation result based on the recommendation score comprises: and sequencing the city feature data from high to low according to the recommendation scores of the city feature data in the candidate set, selecting the city feature data with the preset quantity as recommendation results and outputting the recommendation results.

7. The method of claim 6, wherein said inputting each city feature data in the candidate set to the optimal model, obtaining a corresponding output value comprises:

for each city characteristic data in the candidate set, inputting the city characteristic data into the optimal model for multiple times to obtain multiple output values;

And performing cross-validation optimization on the plurality of output values to obtain optimized output values, and taking the optimized output values as recommendation scores of the city feature data.

8. The method of claim 1, wherein obtaining each type of city feature data as source data comprises:

and cleaning the acquired city characteristic data, wherein the city characteristic data after data cleaning form source data.

9. An urban target recommendation device, comprising:

the first acquisition module is used for acquiring city characteristic data of each type as source data;

the second acquisition module is used for acquiring the type of the target to be recommended;

the recommendation processing module is used for taking the city characteristic data which accords with the target type and accords with the preset recommendation condition in the source data as a candidate set and taking the city characteristic data which accords with the target type and does not accord with the preset recommendation condition in the source data as a sample set; adding an output value mark corresponding to the city feature data into each city feature data in the sample set; training an initial model based on the sample set added with the output value mark to obtain an optimal model; obtaining recommendation scores corresponding to the city feature data in the candidate set based on the optimal model; and outputting a recommendation result based on the recommendation score.

10. The apparatus of claim 9, wherein the first acquisition module acquiring each type of city feature data as source data comprises:

the first acquisition module is used for dividing the area where the city is located into a plurality of area blocks according to a first preset rule; dividing the acquired city characteristic data into area blocks; in the source data, different types of city feature data within the same region block are associated, and/or the same type of city feature data in different region blocks are associated.

11. The apparatus of claim 9, wherein the recommendation processing module trains an initial model based on the sample set with the output value markers added thereto, resulting in an optimal model comprising:

the recommendation processing module is used for extracting a plurality of training sets from the sample set according to a second preset rule, each training set comprises a plurality of pieces of city characteristic data with output value marks, and each training set is different; for each training set, taking part or all of city characteristic data except the training set in the sample set as a verification set corresponding to the training set, training an initial model by using the training set to obtain a training model, verifying the training model by using the verification set, and obtaining an evaluation score of the training model according to a verification result; and taking the training model with the highest evaluation score as an optimal model, and taking a training set corresponding to the training model as an optimal feature combination.

12. The apparatus of claim 11, wherein the recommendation processing module extracting a plurality of training sets from the sample set according to a second preset rule comprises:

the recommendation processing module is used for selecting the city feature data for multiple times in a forward and/or backward traversal mode by adopting a breadth-first search algorithm in the sample set to obtain each training set.

13. The apparatus of claim 11, wherein the recommendation processing module extracting a plurality of training sets from the sample set according to a second preset rule comprises:

the recommendation processing module is used for changing the extracted training set by using a simulated annealing algorithm and/or a random search method to obtain a new training set.

14. The apparatus of claim 11, wherein:

the recommendation processing module obtaining recommendation scores corresponding to the city feature data in the candidate set based on the optimal model comprises the following steps: the recommendation processing module is used for inputting the urban feature data in the candidate set to the optimal model to obtain corresponding output values, and the output value corresponding to each urban feature data is a recommendation score corresponding to the urban feature data;

The recommendation processing module outputting a recommendation result based on the recommendation score comprises: and the recommendation processing module is used for sequencing the urban characteristic data from high to low according to the recommendation scores of the urban characteristic data in the candidate set, selecting the urban characteristic data with the preset quantity as recommendation results and outputting the recommendation results.

15. The apparatus of claim 14, wherein the recommendation processing module inputs each city feature data in the candidate set to the optimal model, obtaining a respective output value comprises:

the recommendation processing module is used for inputting the city characteristic data into the optimal model for a plurality of times for each city characteristic data in the candidate set to obtain a plurality of output values; and performing cross-validation optimization on the plurality of output values to obtain optimized output values, and taking the optimized output values as recommendation scores of the city feature data.

16. The apparatus of claim 9, wherein the first acquisition module acquiring each type of city feature data as source data comprises:

the first acquisition module is used for carrying out data cleaning on the acquired city characteristic data, and the city characteristic data after data cleaning form source data.

17. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the urban objective recommendation method according to any one of claims 1 to 8 when the program is executed.

18. A computer readable medium having stored thereon executable instructions which when executed by a processor cause the processor to perform the urban objective recommendation method according to any one of claims 1 to 8.