Store recommendation method and device, electronic equipment and storage medium

Abstract

The invention belongs to the technical field of data processing, and particularly relates to a store recommendation method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring demand data of a target user and store data of a plurality of stores; inputting the store data into a store recommendation model to obtain recommendation indexes of the plurality of stores, the store recommendation model being obtained by training according to performance data and activity data of the plurality of stores; and based on the recommendation indexes of the plurality of stores and the demand data, determining a target recommendation store in the plurality of stores. The method is used for achieving the effects of improving the accuracy of store recommendation based on the user demand, the geographic position of the store and the store image, and further providing the optimal store recommendation for the user.

Description

Technical Field

[0001] This application belongs to the field of data processing technology, specifically relating to a store recommendation method, apparatus, electronic device, and storage medium. Background Technology

[0002] With the widespread application of big data in the recommendation field, more and more industries are providing personalized recommendation services to users by analyzing and processing relevant data. For example, in the field of store recommendations, existing technologies typically capture users' interests and preferences based on their historical behavioral data, and then provide users with stores that they are interested in.

[0003] In the process of matching users with stores, existing technologies may also consider the store's geographical location and the user's current needs to match the user with the target store.

[0004] However, existing technologies fail to fully consider the causal relationship between geographic location data, store characteristics, and user needs when processing real-time data and responding to user needs, resulting in low accuracy of recommendation results and thus failing to provide users with the optimal user experience and store recommendation services. Summary of the Invention

[0005] This application provides a store recommendation method, apparatus, electronic device, and storage medium to solve the problem that the prior art does not consider the causal relationship between store characteristics and user needs, resulting in low accuracy of recommendation results.

[0006] On the one hand, this application provides a store recommendation method, including:

[0007] Obtain target user demand data and store data from multiple stores;

[0008] The store data is input into the store recommendation model to obtain the recommendation index of the multiple stores. The store recommendation model is trained based on the performance data and activity data of the multiple stores.

[0009] Based on the recommendation metrics of the multiple stores and the demand data, a target recommended store is determined among the multiple stores.

[0010] In one possible implementation, determining the target recommended store among the plurality of stores based on the recommendation metrics of the plurality of stores and the demand data includes:

[0011] Based on the recommendation metrics of the multiple stores and the location data in the demand data, a first recommendation coefficient for the multiple stores is determined;

[0012] Determine whether the first recommendation coefficient reaches the preset recommendation coefficient;

[0013] If the first recommendation coefficient reaches the preset recommendation coefficient, at least one target recommended store is determined based on the first recommendation coefficient;

[0014] If the first recommendation coefficient does not reach the preset recommendation coefficient, the demand data and the store data corresponding to the recommendation index are matched to obtain the second recommendation coefficient of the multiple stores;

[0015] Based on the second recommendation coefficient, a target recommended store is determined among the plurality of stores.

[0016] In one possible implementation, after matching the demand data with the store data corresponding to the recommendation index to obtain the second recommendation coefficient for the plurality of stores, the method further includes:

[0017] Based on the business strategies of the multiple stores and the recommendation indicators, a third recommendation coefficient is determined for the multiple stores, wherein the business strategies include the current marketing strategies of the corresponding stores;

[0018] The recommendation weights of the multiple stores are determined based on the second recommendation coefficient and the third recommendation coefficient.

[0019] Based on the recommendation weight, a target recommended store is determined among the plurality of stores.

[0020] In one possible implementation, before inputting the store data into the store recommendation model to obtain the recommendation metrics for the plurality of stores, the method further includes:

[0021] Obtain historical store datasets and classification thresholds. The historical store datasets include performance datasets and activity datasets. Each historical store data in the historical store dataset carries a historical data category, and the historical data category corresponds one-to-one with the recommendation metric.

[0022] Determine the first feature value of each historical store data in the historical store dataset;

[0023] When the maximum first feature value reaches the classification threshold, the historical data category corresponding to the maximum first feature value is determined as the first classification feature;

[0024] The historical dataset is divided into multiple subsets of first historical store data using the first classification feature;

[0025] For any subset of the plurality of first historical store data subsets, the recommendation index corresponding to the first historical store data subset is determined based on the recommendation index corresponding to the historical data type of each historical store data in the first historical store data subset.

[0026] The store recommendation model is constructed based on the historical store data corresponding to the first classification feature, the first subset of historical store data, and the recommendation index.

[0027] In one possible implementation, after dividing the historical dataset into multiple first historical store data subsets using the first classification feature, the method further includes:

[0028] For any subset of the plurality of first historical store data subsets, determine whether the first historical store data subset meets the preset termination classification condition. The termination classification condition includes that the data categories corresponding to the historical store data in the first historical store data subset are consistent.

[0029] If the first subset of historical store data does not meet the termination classification condition, determine the second feature value of each historical store data in the first subset of historical store data, and determine the second classification feature based on the second feature value.

[0030] The first historical store data subset is classified according to the second classification feature to obtain at least one second historical store data subset.

[0031] In one possible implementation, acquiring the target user's demand data and the store data of multiple stores includes:

[0032] Based on the target user's actions on the user interface, determine the target user's demand data;

[0033] Based on the demand data, determine the store data of multiple stores within a preset time period and preset range.

[0034] In one possible implementation, after constructing the store recommendation model based on the historical store data corresponding to the first classification feature, the first subset of historical store data, and the recommendation metric, the method further includes:

[0035] Monitor the performance and activity data of the aforementioned multiple stores;

[0036] The store recommendation model is optimized based on the performance data and the activity data.

[0037] Secondly, embodiments of this application provide a store recommendation device, comprising:

[0038] The acquisition module is used to acquire the target user's demand data and store data from multiple stores;

[0039] The processing module is used to input the store data into the store recommendation model to obtain the recommendation index of the multiple stores;

[0040] The determination module is used to determine the target recommended store among the multiple stores based on the recommendation indicators of the multiple stores and the demand data.

[0041] In one possible implementation, the device further includes: a matching module;

[0042] The processing module is also used to determine whether the first recommendation coefficient reaches the preset recommendation coefficient;

[0043] The determining module is further configured to determine at least one target recommended store based on the first recommendation coefficient when the first recommendation coefficient reaches the preset recommendation coefficient.

[0044] The matching module is used to match the demand data and the store data corresponding to the recommendation index to obtain the second recommendation coefficient of the multiple stores when the first recommendation coefficient does not reach the preset recommendation coefficient.

[0045] The determining module is further configured to determine a target recommended store among the plurality of stores based on the second recommendation coefficient.

[0046] In one possible implementation, the determining module is further configured to determine a third recommendation coefficient for the plurality of stores based on the business strategies of the plurality of stores and the recommendation indicators, wherein the business strategies include the current marketing strategies of the corresponding stores;

[0047] The determining module is further configured to determine the recommendation weight of the plurality of stores based on the second recommendation coefficient and the third recommendation coefficient;

[0048] The determining module is further configured to determine the target recommended store among the plurality of stores based on the recommendation weight.

[0049] In one possible implementation, the device further includes a classification module and a construction module;

[0050] The acquisition module is also used to acquire historical store datasets and classification thresholds;

[0051] The determining module is further configured to determine the first feature value of each historical store data in the historical store dataset;

[0052] The determining module is further configured to determine the historical data category corresponding to the maximum first feature value as the first classification feature when the maximum first feature value reaches the classification threshold;

[0053] The classification module is used to divide the historical dataset into multiple first historical store data subsets using the first classification feature;

[0054] The determining module is further configured to, for any first historical store data subset in the plurality of first historical store data subsets, determine the recommendation index corresponding to the first historical store data subset based on the recommendation index corresponding to the historical data type of each historical store data in the first historical store data subset;

[0055] The construction module is used to construct the store recommendation model based on the historical store data corresponding to the first classification feature, the first subset of historical store data, and the recommendation index.

[0056] In one possible implementation, the processing module is further configured to determine whether any first historical store data subset in the plurality of first historical store data subsets meets a preset termination classification condition.

[0057] The determining module is further configured to determine the second feature value of each historical store data in the first historical store data subset when the first historical store data subset does not meet the termination classification condition, and determine the second classification feature based on the second feature value.

[0058] The classification module is further configured to classify the first historical store data subset according to the second classification feature to obtain at least one second historical store data subset.

[0059] In one possible implementation, the determining module is further configured to determine the target user's demand data based on the target user's operations on the user interface;

[0060] The determining module is also used to determine store data for multiple stores within a preset time period and preset range based on the demand data.

[0061] In one possible implementation, the device further includes: a monitoring device;

[0062] The monitoring device is used to monitor the performance and activity data of the multiple stores;

[0063] The processing module is also used to optimize the store recommendation model based on the performance data and the activity data.

[0064] Thirdly, this application provides a store recommendation device, comprising:

[0065] Memory;

[0066] processor;

[0067] The memory stores computer-executed instructions;

[0068] The processor executes computer execution instructions stored in the memory to implement the store recommendation method as described in the first aspect and various possible implementations of the first aspect above.

[0069] Fourthly, this application provides a computer storage medium storing computer execution instructions, which are executed by a processor to implement the store recommendation method as described in the first aspect and various possible implementations thereof.

[0070] This application provides a store recommendation method, apparatus, electronic device, and storage medium. By acquiring target user demand data and store data from multiple stores, and employing a store recommendation model to process the store data, recommendation indicators corresponding to the multiple stores are obtained. Then, based on the target user's demand data, the target recommended store is determined. This method combines user demand and store data, and based on the store recommendation model, determines the correlation between user demand and store characteristics, improving the accuracy of store recommendations and thus providing users with a better store recommendation service. Attached Figure Description

[0071] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0072] Figure 1 A flowchart illustrating a store recommendation method provided in this application. Figure 1 ;

[0073] Figure 2 A schematic diagram of the structure of a store recommendation model provided in this application;

[0074] Figure 3 A flowchart illustrating a store recommendation method provided in this application. Figure 2 ;

[0075] Figure 4 A flowchart illustrating a store recommendation method provided in this application. Figure 3 ;

[0076] Figure 5 A schematic diagram of the structure of a store recommendation device provided in this application;

[0077] Figure 6 This is a structural diagram of a store recommendation device provided in this application.

[0078] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0079] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0080] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented, for example, in orders other than those illustrated or described herein.

[0081] In this application, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0082] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with relevant rules and standards, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0083] With the application of big data and artificial intelligence technologies in the recommendation field, existing recommendation technologies are increasingly focusing on personalized recommendation services. For example, in the field of store recommendations, by collecting and analyzing interactive data such as users' purchase history, browsing records, and search behavior, users' consumption habits and preferences are determined, and stores that users are interested in are recommended based on these habits and preferences.

[0084] However, existing technologies fail to fully consider the causal relationship between geographic location data, store characteristics, and user needs when processing interactive data and responding to user needs, which reduces the accuracy of recommendation results and thus fails to provide users with the optimal user experience and store recommendation services.

[0085] To address the aforementioned issues, this application provides a store recommendation method. This method processes store data using a store recommendation model, determines recommendation metrics for multiple stores based on the correlation between historical demand data of multiple users and store data, and then identifies target recommended stores based on these metrics and the target user's demand data, thus satisfying the user's personalized needs. Furthermore, by determining the correlation between the target user's demand data and store data, this method not only improves the accuracy of store recommendations but also incorporates store characteristics to provide users with highly tailored store recommendation services.

[0086] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0087] Figure 1 The process of a store recommendation method provided in this application embodiment Figure 1 .like Figure 1 As shown, this embodiment provides a store recommendation method, including:

[0088] S101: Obtain target user demand data and store data from multiple stores.

[0089] The demand data includes the product type and store location data set by the target user. The store data includes store performance data and store activity data. Store performance data may include sales data, customer traffic and profit. Activity data includes the duration, number of participating customers and promotional volume of promotional activities and advertising campaigns conducted by the store.

[0090] Optionally, multiple stores within a preset distance can be identified based on the location information in the demand data, and store data for each store can be obtained. For example, sales data, customer traffic, profit, and other performance-related data can be extracted from existing stores' POS (point of sale) systems or enterprise databases. Additionally, demand data can be submitted by customers to a store recommendation system.

[0091] S102: Input the store data into the store recommendation model to obtain the recommendation index of the multiple stores.

[0092] The store recommendation model is trained based on the performance and activity data of the multiple stores, and the store recommendation model includes store recommendation indicators corresponding to the performance and activity data.

[0093] Optionally, multiple store data are input into the store recommendation model. Based on the data type and feature value of the multiple store data, the optimal classification feature is determined, and the multiple store data are divided according to the classification feature to obtain the classification result. Then, the recommendation index of the corresponding store is determined based on the classification result. Figure 2 This is a structural diagram of a store recommendation model. Figure 2 The classification results shown use the data types of performance and activity data within the store data as classification nodes to categorize the store data, resulting in classification results. Corresponding recommendation indicators are then determined based on these classification results. Feature values ​​are used to characterize the importance of the corresponding store data; data types can include, for example, sales revenue, customer traffic, and activity participation.

[0094] S103: Based on the recommendation indicators of the multiple stores and the demand data, determine the target recommended store among the multiple stores.

[0095] Optionally, by matching demand data and store data, target recommended stores are determined based on the matching degree and the recommendation metrics corresponding to the store data. For example, at least one store corresponding to store data that reaches a preset matching degree is identified as a store to be recommended, and the target recommended store is determined based on the recommendation metrics corresponding to the store to be recommended.

[0096] This embodiment provides a store recommendation method that acquires target user demand data and store data from multiple stores. It then processes the store data using a store recommendation model to obtain recommendation indicators for each store. Based on the target user's demand data, it determines the target recommended store. This method combines user demand and store data, and based on the store recommendation model, establishes the correlation between user demand and store characteristics, improving the accuracy of store recommendations and providing users with a superior store recommendation service.

[0097] Figure 3 The process of a store recommendation method provided in this application embodiment Figure 2 .like Figure 3 As shown, in this embodiment... Figure 1 Based on the embodiments, a store recommendation method is described in detail. This embodiment illustrates a store recommendation method, which includes:

[0098] S301: Obtain target user demand data and store data from multiple stores.

[0099] Optionally, based on the target user's actions on the user interface, the target user's demand data is determined, and then based on the demand data, store data for multiple stores within a preset time period and preset range is determined.

[0100] Optionally, the preset time period can be set by the target user, and the preset range can be determined based on the target user's location information.

[0101] S302: Input the store data into the store recommendation model to obtain the recommendation index of the multiple stores.

[0102] The explanation of step S302 is similar to that of step S102 above, and will not be repeated here.

[0103] S303: Based on the recommendation indicators of the multiple stores and the location data in the demand data, determine the first recommendation coefficient of the multiple stores.

[0104] The first recommendation coefficient is used to characterize the recommendation weight of each store.

[0105] Optionally, for multiple stores' recommendation metrics, it is determined whether the recommendation metric meets a preset recommendation metric. Stores meeting the preset recommendation metric are then identified as stores to be recommended. A first recommendation coefficient for each store is determined based on the location data in the demand data and the location data of the stores to be recommended. For example, the distance difference between the location data in the demand data and the location data of the stores to be recommended is determined, and the first recommendation coefficient is determined based on this distance difference and the recommendation metric of the stores to be recommended. The smaller the distance difference and the higher the recommendation metric, the larger the corresponding first recommendation coefficient. In this method, the first recommendation coefficient is jointly determined by the location data in the target demand and the recommendation metrics of multiple stores. This considers both the user's demand for store location and the service advantages of the stores themselves, which is conducive to providing users with better services.

[0106] S304: Determine whether the first recommendation coefficient has reached the preset recommendation coefficient.

[0107] Optionally, based on the stores selected by the user according to historical recommendations, the system can capture the user's satisfaction with the stores to be recommended, determined according to different recommendation indicators. A preset recommendation coefficient can then be determined based on this satisfaction level. During the store recommendation process, this preset recommendation coefficient can be used to determine whether the stores to be recommended meet the user's satisfaction level. For example, if the first store's first recommendation coefficient reaches the preset recommendation coefficient, it means that the first store meets the user's satisfaction level in terms of distance, and it can be identified as a store to be recommended.

[0108] S305: If the first recommendation coefficient reaches the preset recommendation coefficient, at least one target recommended store shall be determined based on the first recommendation coefficient.

[0109] Optionally, if multiple stores achieve a preset recommendation coefficient, the stores are ranked according to this coefficient, and the target recommended store is determined based on this ranking. For example, multiple stores can be recommended to the target user in the order of their ranking. Alternatively, the top three stores can be selected as the target recommended stores and recommended to the target user.

[0110] S306: If the first recommendation coefficient does not reach the preset recommendation coefficient, the demand data and the store data corresponding to the recommendation index are matched to obtain the second recommendation coefficient of the multiple stores.

[0111] Optionally, if the first recommendation coefficient does not reach the preset recommendation coefficient, it indicates that the store corresponding to the first recommendation coefficient does not meet user satisfaction. Therefore, the demand data is matched with the classification results output by the store recommendation model, and a second recommendation coefficient for the corresponding store is determined based on the matching degree and the corresponding store's recommendation index. The higher the matching degree and the higher the recommendation index, the larger the recommendation coefficient for the store. The demand data includes user needs such as product type, location data, product manufacturer, and current promotional discounts. This method comprehensively considers other user needs when the stores to be recommended based on location data do not meet user satisfaction, in order to improve the accuracy of store recommendations and provide users with satisfactory recommendation services.

[0112] It should be noted that the target user's demand data can be submitted by the user. If the user does not submit sufficient demand data, the collected user demand dataset can be analyzed to determine the target user's common demands for the submitted products. These common demands can then be identified as the target user's demand data, in order to provide users with more personalized recommendation services.

[0113] S307: Based on the second recommendation coefficient, determine the target recommended store among the plurality of stores.

[0114] Optionally, after determining the second recommendation coefficient for multiple stores, similar to step S305 above, the multiple stores can be ranked according to the second recommendation coefficient, and the target recommended store can be determined based on the ranking.

[0115] This application provides a store recommendation method that determines a first recommendation coefficient for multiple stores based on location data and recommendation indicators of multiple stores in demand data. If the first recommendation coefficient does not reach a preset recommendation coefficient, the method considers all the needs of the target user, matches the target user's demand information with the store data, and determines a second recommendation coefficient based on the matching degree and the corresponding store's recommendation indicators. Finally, the method determines the target recommended store based on the second recommendation coefficient. This method first considers the location distance between the store and the target user and provides store recommendation services based on the first recommendation coefficient. If the first recommendation coefficient, i.e., the location distance, does not meet the user's needs, it comprehensively considers the target user's demand data to provide a more accurate recommendation service, improving user satisfaction. Furthermore, this method comprehensively considers other needs related to the target user's current needs, providing a more comprehensive and user-friendly store recommendation service.

[0116] In one possible implementation, if the first recommendation coefficient does not reach the preset recommendation coefficient, the demand data and the store data corresponding to the recommendation indicators are matched to obtain the second recommendation coefficient for the multiple stores. Furthermore, the business strategies of the stores can be combined to recommend more suitable stores to users, and targeted traffic can be directed to stores employing different business strategies. Figure 1-3 Based on the examples, a method for determining target recommended stores by combining store business strategies is described in detail. This method specifically includes:

[0117] Based on the business strategies of the multiple stores and the recommendation metrics, a third recommendation coefficient is determined for the multiple stores, wherein the business strategies include the current marketing strategies of the corresponding stores.

[0118] The business strategy includes factors such as the store's current marketing strategy, target customer distribution, market competitive advantages, and store operating time.

[0119] Specifically, the business strategies of stores are matched with the needs data of target users to obtain the matching degree between target users and corresponding stores. Then, based on the matching degree and the recommendation indicators of the corresponding stores, a third recommendation coefficient for the corresponding stores is determined. For example, the business strategies of the first store are matched with the needs data of target users. Based on the obtained matching degree and the recommendation indicators of the first store, the recommendation coefficient of the first store is determined. The higher the matching degree and recommendation indicators of the first store, the larger the recommendation coefficient.

[0120] The recommendation weights of the multiple stores are determined based on the second recommendation coefficient and the third recommendation coefficient.

[0121] Optionally, the second recommendation coefficient is a store recommendation coefficient that takes into account the target user's demand data and the store's recommendation indicators. The third recommendation coefficient is obtained by comprehensively considering the target user's demand and the store's business strategy. The store's recommendation indicators are determined based on the second and third recommendation coefficients, which further considers the relationship between user demand and store characteristics, improves the fit between users and target recommended stores, and thus provides users with more accurate store recommendation services.

[0122] It should be noted that the recommendation weight can be determined by the average of the second recommendation coefficient and the third recommendation coefficient, or the proportion of the second recommendation coefficient and the third recommendation coefficient in the recommendation weight can be preset, and the weight of the second recommendation coefficient and the third recommendation coefficient can be determined according to their proportions, and the weights can be summed to obtain the recommendation weight of the corresponding store.

[0123] Based on the recommendation weight, a target recommended store is determined among the plurality of stores.

[0124] Optionally, stores can be sorted according to their recommendation weights, and target recommended stores can be determined based on the sorting results. Alternatively, target recommended stores can be determined based on preset recommendation weights. For example, if the recommendation weight of the first store reaches the preset recommendation weight, the first store can be identified as the target recommended store.

[0125] This application provides a store recommendation method that, after considering the needs of target users and store data, and determining a second recommendation coefficient for each store, a third recommendation coefficient is determined by combining the store's business strategy. Based on this third recommendation coefficient and the corresponding second recommendation coefficient, a recommendation weight is determined for each store. Finally, based on this recommendation weight, it is determined whether the store is a target recommended store, thus improving the accuracy of store recommendations. Furthermore, this method fully considers user needs and store characteristics, and incorporates the store's current business strategy, providing the most suitable store recommendation service for target users.

[0126] Figure 4 The process of a store recommendation method provided in this embodiment Figure 3 .like Figure 4 As shown, in this embodiment... Figure 1 , 3 Based on the embodiments, a training method for a store recommendation model is described in detail. This embodiment illustrates a store recommendation method, including:

[0127] S401: Obtain historical store datasets and classification thresholds.

[0128] The historical store dataset includes a performance dataset and an activity dataset. Each historical store data point in the historical store dataset carries historical data categories and recommendation metrics. The classification threshold is used to evaluate whether to classify the store dataset using the current features. For example, if the feature value corresponding to sales revenue is greater than the classification threshold, then sales revenue is determined as the classification feature, and the store dataset is classified.

[0129] Optionally, after obtaining the historical store dataset, data cleaning and preprocessing can be performed on the dataset to ensure data consistency and accuracy, facilitating the training and analysis of the store recommendation model. Data cleaning includes removing duplicate data and handling missing and outlier values, while data preprocessing includes standardization and normalization.

[0130] S402: Determine the first feature value of each historical store data in the historical store dataset.

[0131] Optionally, a first feature value for each historical store data point is determined based on its importance within the historical store dataset. This first feature value can be the information gain, which can then be used to determine the information gain of the historical store data within the historical data category set based on its information entropy and conditional entropy.

[0132] S403: When the maximum first feature value reaches the classification threshold, the historical data category corresponding to the maximum first feature value is determined as the first classification feature.

[0133] Specifically, based on the classification threshold and the first feature value corresponding to the historical store data, the optimal classification feature is determined and designated as the first classification feature. This first classification feature is the historical data feature corresponding to the historical store data.

[0134] Optionally, based on the feature values ​​corresponding to each historical store data in the historical store dataset, the largest feature value is determined in the historical store dataset, and it is determined whether this feature value is greater than or equal to a classification threshold. If the feature value is greater than or equal to the classification threshold, the historical store data corresponding to this feature value is determined, and the historical data category corresponding to this historical store data is determined as the first classification feature. For example, the classification threshold can be an information gain threshold, and the feature value is information gain. During the classification of the historical store dataset, it is determined whether the information gain of each historical store data is greater than or equal to the information gain threshold. If the information gain is greater than or equal to the information gain threshold, the data type corresponding to that information gain is used as a classification node to divide the historical store dataset; if the information gain is less than the information gain threshold, the data type corresponding to that information gain cannot be used as a classification node.

[0135] S404: The historical dataset is divided into multiple first historical store data subsets using the first classification feature.

[0136] Optionally, the first classification feature can be used as a classification node to divide the historical store dataset, resulting in multiple first historical store data subsets. Each first historical store data subset no longer contains historical store data corresponding to the first classification feature.

[0137] S405: For any first historical store data subset in the plurality of first historical store data subsets, determine the recommendation index corresponding to the first historical store data subset based on the recommendation index corresponding to the historical data type of each historical store data in the first historical store data subset.

[0138] Optionally, after the classification operation of the historical store dataset is completed and multiple first historical store data subsets are obtained, the recommendation metrics corresponding to the historical data types of each historical store data subset are comprehensively considered, and the recommendation metrics for each first historical store data subset are determined. For example, if the historical data types and corresponding recommendation metrics in the first historical store data subsets include: (sales volume, high), (customer traffic, high), and (participation in new product discounts, medium), then based on the historical data types and corresponding recommendation metrics, the recommendation metric for the corresponding first historical store data subset can be determined as a high recommendation metric.

[0139] Optionally, in determining the recommended metrics corresponding to the first subset of historical store data, the recommended metric with the highest number of different recommended metrics can be selected as the recommended metric for the first subset of historical store data. Alternatively, if the number of each recommended metric corresponding to the first subset of historical store data is consistent, the recommended metric for the first subset of historical store data can also be determined based on the number of each recommended metric and its preset weight.

[0140] S406: Based on the historical store data corresponding to the first classification feature, the first subset of historical store data, and the recommendation index, construct the store recommendation model.

[0141] Optionally, the first classification feature is used as the classification node, and a store recommendation model is constructed based on the historical data type and recommendation metrics corresponding to each historical store in the first historical store data subset. For example... Figure 2 The classification nodes include data types such as product type, sales volume, service satisfaction, customer traffic, and participation in new product discount activities. Based on the above data types, a decision tree model is constructed, and the recommendation index for each subset obtained from the classification is determined according to the recommendation index corresponding to each data type.

[0142] Optionally, performance and activity data from multiple stores can be monitored, and the store recommendation model can be optimized based on this performance and activity data.

[0143] In one possible implementation, after dividing the historical dataset into multiple first historical store data subsets using the first classification feature, a detailed description is provided of a method for determining whether to continue classifying the first historical store data subsets based on the termination classification condition judgment result. This method includes:

[0144] For any subset of the plurality of first historical store data subsets, determine whether the first historical store data subset meets the preset termination classification condition.

[0145] The termination classification condition includes that the data categories corresponding to the historical store data in the first historical store data subset are consistent.

[0146] Optionally, it can be determined whether the historical data categories in the first historical store data subset are consistent. If the historical data categories of the historical store data in the first historical store data subset are consistent, it can be determined that the first historical store data subset has reached the preset termination classification condition. If the historical data categories of the historical store data in the first historical store data subset are inconsistent, it can be determined that the first historical store data subset has not reached the preset termination classification condition.

[0147] If the first subset of historical store data does not meet the termination classification condition, the second feature value of each historical store data in the first subset of historical store data is determined, and the second classification feature is determined based on the second feature value.

[0148] Optionally, if the first subset of historical store data does not meet the termination classification criteria, a second feature value is determined for each historical store data point based on its importance to the first subset. The historical data category corresponding to the second feature value that reaches the classification threshold is then identified as the second classification feature. For example, the information gain corresponding to the historical data type of each historical store data point in the first subset is calculated, and it is determined whether the information gain reaches the information gain threshold. If the information gain reaches the information gain threshold, the historical data type corresponding to that information gain is identified as the second classification feature.

[0149] The first historical store data subset is classified according to the second classification feature to obtain at least one second historical store data subset.

[0150] Optionally, the second classification feature can be used as a new classification node, and the first historical store data subset can be classified based on this classification node to obtain at least one second historical store data subset. After classification, a store recommendation model is constructed based on the recommendation metrics corresponding to each classification node and each historical data category.

[0151] This application provides a store recommendation method that, by acquiring a historical store dataset and a preset classification threshold, determines the feature values ​​of each historical store data point in the historical store dataset. Based on these feature values ​​and the classification threshold, it determines whether to classify the historical store dataset using the historical data type corresponding to the feature value as the classification feature. After classifying the historical store dataset using this historical data type to obtain a first subset of historical store data, it determines the first subset of historical store data that needs further classification based on a termination classification condition. Then, it divides the first subset of historical store data according to the feature values ​​and historical data types corresponding to each historical store data point, iterating this classification operation until all historical store data subsets reach the termination classification condition. Finally, it determines the recommendation indicators corresponding to each historical store data subset based on the recommendation indicators corresponding to each historical data type. This method constructs a store recommendation model based on historical performance data and historical activity data of stores, not only obtaining the relationship between store characteristics and user needs but also determining recommendation indicators for different store characteristics based on this relationship, thus improving the accuracy of the store recommendation model.

[0152] Figure 5 This is a structural schematic diagram of a store recommendation device provided in this application. Figure 5 As shown, this application provides a store recommendation device, the store recommendation device 50 including:

[0153] The acquisition module 501 is used to acquire the target user's demand data and the store data of multiple stores;

[0154] The processing module 502 is used to input the store data into the store recommendation model to obtain the recommendation index of the multiple stores;

[0155] The determination module 503 is used to determine the target recommended store among the multiple stores based on the recommendation indicators of the multiple stores and the demand data.

[0156] In one possible implementation, the device further includes: a matching module 504;

[0157] The processing module 502 is further configured to determine whether the first recommendation coefficient reaches a preset recommendation coefficient;

[0158] The determining module 503 is further configured to determine at least one target recommended store based on the first recommendation coefficient when the first recommendation coefficient reaches the preset recommendation coefficient.

[0159] The matching module 504 is used to match the demand data and the store data corresponding to the recommendation index to obtain the second recommendation coefficient of the multiple stores when the first recommendation coefficient does not reach the preset recommendation coefficient.

[0160] The determining module 503 is further configured to determine a target recommended store among the plurality of stores based on the second recommendation coefficient.

[0161] In one possible implementation, the determining module 503 is further configured to determine a third recommendation coefficient for the plurality of stores based on the business strategies of the plurality of stores and the recommendation indicators, wherein the business strategies include the current marketing strategies of the corresponding stores.

[0162] The determining module 503 is further configured to determine the recommendation weight of the plurality of stores based on the second recommendation coefficient and the third recommendation coefficient;

[0163] The determining module 503 is further configured to determine the target recommended store among the plurality of stores based on the recommendation weight.

[0164] In one possible implementation, the device further includes: a classification module 505 and a construction module 506;

[0165] The acquisition module 501 is also used to acquire historical store datasets and classification thresholds;

[0166] The determining module 503 is further configured to determine the first feature value of each historical store data in the historical store dataset;

[0167] The determining module 503 is further configured to determine the historical data category corresponding to the maximum first feature value as the first classification feature when the maximum first feature value reaches the classification threshold.

[0168] The classification module 505 is used to divide the historical dataset into multiple first historical store data subsets using the first classification feature;

[0169] The determining module 503 is further configured to, for any first historical store data subset in the plurality of first historical store data subsets, determine the recommendation index corresponding to the first historical store data subset based on the recommendation index corresponding to the historical data type of each historical store data in the first historical store data subset;

[0170] The construction module 506 is used to construct the store recommendation model based on the historical store data corresponding to the first classification feature, the first subset of historical store data, and the recommendation index.

[0171] In one possible implementation, the processing module 502 is further configured to determine whether the first historical store data subset has reached a preset termination classification condition for any first historical store data subset among the plurality of first historical store data subsets.

[0172] The determining module 503 is further configured to determine the second feature value of each historical store data in the first historical store data subset when the first historical store data subset does not meet the termination classification condition, and determine the second classification feature based on the second feature value.

[0173] The classification module 505 is further configured to classify the first historical store data subset according to the second classification feature to obtain at least one second historical store data subset.

[0174] In one possible implementation, the determining module 503 is further configured to determine the target user's demand data based on the target user's operations on the user interface;

[0175] The determining module 503 is also used to determine store data of multiple stores within a preset time period and preset range based on the demand data.

[0176] In one possible implementation, the device further includes: a monitoring device 507;

[0177] The monitoring device 507 is used to monitor the performance data and activity data of the multiple stores;

[0178] The processing module 502 is further configured to optimize the store recommendation model based on the performance data and the activity data.

[0179] This embodiment provides a store recommendation device that can execute the method provided in the above-described method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0180] Figure 6 This is a structural diagram of a store recommendation device provided in this application. Figure 6 As shown, this application provides a store recommendation device 60, which includes: a receiver 601, a transmitter 602, a processor 603, and a memory 604.

[0181] Receiver 601 is used to receive instructions and data;

[0182] Transmitter 602 is used to send commands and data;

[0183] Memory 604 is used to store instructions executed by the computer;

[0184] The processor 603 is used to execute computer execution instructions stored in the memory 604 to implement the various steps of the store recommendation method in the above embodiments. For details, please refer to the relevant description in the foregoing embodiments of the store recommendation method.

[0185] Alternatively, the memory 604 can be either standalone or integrated with the processor 403.

[0186] When the memory 604 is set up independently, the electronic device also includes a bus for connecting the memory 604 and the processor 603.

[0187] This application also provides a computer storage medium storing computer execution instructions, which, when executed by a processor, implement a store recommendation method as described above by a store recommendation device.

[0188] It will be understood by those skilled in the art that all or some of the steps, systems, or apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software may be distributed on a computer-readable medium, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0189] The technical solutions of this application have been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it is readily understood by those skilled in the art that the scope of protection of this application is obviously not limited to these specific embodiments. The above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A store recommendation method, characterized in that, include: Obtain target user demand data and store data from multiple stores; The store data is input into the store recommendation model to obtain the recommendation index of the multiple stores. The store recommendation model is trained based on the performance data and activity data of the multiple stores. Based on the recommendation metrics of the multiple stores and the demand data, a target recommended store is determined among the multiple stores.

2. The method according to claim 1, characterized in that, The step of determining the target recommended store among the multiple stores based on the recommendation metrics of the multiple stores and the demand data includes: Based on the recommendation metrics of the multiple stores and the location data in the demand data, a first recommendation coefficient for the multiple stores is determined; Determine whether the first recommendation coefficient reaches the preset recommendation coefficient; If the first recommendation coefficient reaches the preset recommendation coefficient, at least one target recommended store is determined based on the first recommendation coefficient; If the first recommendation coefficient does not reach the preset recommendation coefficient, the demand data and the store data corresponding to the recommendation index are matched to obtain the second recommendation coefficient of the multiple stores; Based on the second recommendation coefficient, a target recommended store is determined among the plurality of stores.

3. The method according to claim 2, characterized in that, After matching the demand data with the store data corresponding to the recommendation indicators to obtain the second recommendation coefficient for the multiple stores, the method further includes: Based on the business strategies of the multiple stores and the recommendation indicators, a third recommendation coefficient is determined for the multiple stores, wherein the business strategies include the current marketing strategies of the corresponding stores; The recommendation weights of the multiple stores are determined based on the second recommendation coefficient and the third recommendation coefficient. Based on the recommendation weight, a target recommended store is determined among the plurality of stores.

4. The method according to claim 1, characterized in that, Before inputting the store data into the store recommendation model to obtain the recommendation metrics for the multiple stores, the method further includes: Obtain historical store datasets and classification thresholds. The historical store datasets include performance datasets and activity datasets. Each historical store data in the historical store dataset carries a historical data category, and the historical data category corresponds one-to-one with the recommendation metric. Determine the first feature value of each historical store data in the historical store dataset; When the maximum first feature value reaches the classification threshold, the historical data category corresponding to the maximum first feature value is determined as the first classification feature; The historical dataset is divided into multiple subsets of first historical store data using the first classification feature; For any subset of the plurality of first historical store data subsets, the recommendation index corresponding to the first historical store data subset is determined based on the recommendation index corresponding to the historical data type of each historical store data in the first historical store data subset. The store recommendation model is constructed based on the historical store data corresponding to the first classification feature, the first subset of historical store data, and the recommendation index.

5. The method according to claim 4, characterized in that, After dividing the historical dataset into multiple subsets of first historical store data using the first classification feature, the method further includes: For any subset of the plurality of first historical store data subsets, determine whether the first historical store data subset meets the preset termination classification condition. The termination classification condition includes that the data categories corresponding to the historical store data in the first historical store data subset are consistent. If the first subset of historical store data does not meet the termination classification condition, determine the second feature value of each historical store data in the first subset of historical store data, and determine the second classification feature based on the second feature value. The first historical store data subset is classified according to the second classification feature to obtain at least one second historical store data subset.

6. The method according to claim 1, characterized in that, The acquisition of target user demand data and store data from multiple stores includes: Based on the target user's actions on the user interface, determine the target user's demand data; Based on the demand data, determine the store data of multiple stores within a preset time period and preset range.

7. The method according to claim 4, characterized in that, After constructing the store recommendation model based on the historical store data corresponding to the first classification feature, the first subset of historical store data, and the recommendation metric, the method further includes: Monitor the performance and activity data of the aforementioned multiple stores; The store recommendation model is optimized based on the performance data and the activity data.

8. A store recommendation device, characterized in that, include: The acquisition module is used to acquire the target user's demand data and store data from multiple stores; The processing module is used to input the store data into the store recommendation model to obtain the recommendation index of the multiple stores; The determination module is used to determine the target recommended store among the multiple stores based on the recommendation indicators of the multiple stores and the demand data.

9. A store recommendation device, characterized in that, include: Memory; processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement a store recommendation method as described in any one of claims 1-7.

10. A computer storage medium, characterized in that, The computer storage medium stores computer execution instructions, which, when executed by a processor, are used to implement a store recommendation method as described in any one of claims 1-7.

Images