A method, device, equipment and storage medium for selecting goods for sale in a store

Abstract

The present application relates to a method, device, equipment and storage medium for selecting goods to be put on the shelves of stores. The main technical scheme includes: by obtaining the historical sales information of the pre-selected stores and the geo-tag information of all buildings within the preset range, merging the geo-tag categories according to the functional types of the geo-tags in the geo-tag categories, obtaining the effective tag categories, determining the standardized tag distance value between each pre-selected store and the geo-tag in the effective tag category and the standardized tag quantity value of the geo-tag in the effective tag category of the pre-selected store according to the number of geo-tags corresponding to the pre-selected stores and the average distance between each geo-tag and the pre-selected stores, determining the classification information of the pre-selected stores according to the standardized distance value and the standardized tag quantity value, determining the goods to be put on the shelves of the target stores according to the classification information and historical sales information of the pre-selected stores, optimizing the goods to be put on the shelves of stores with poor sales performance, recommending goods suitable for store sales, and increasing store sales.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method, device, equipment and storage medium for selecting goods for sale in a store. Background Art

[0002] With the continuous innovation of consumption patterns, my country's retail industry is still in a growth trend. The retail industry is mainly divided into online and offline. As the last link of the entire sales channel, offline terminals have the characteristics of thousands of stores with thousands of faces, and their own convenience is difficult to replace by other channels. Brands can improve investment efficiency by optimizing the company's display and other marketing strategies, and at the same time assist channel operators to optimize various aspects of sales, so that a large number of convenience stores can choose products suitable for their sales on limited shelf space, reduce trial and error costs, and increase sales.

[0003] At present, when the market segments stores, it is usually based on qualitative (staff evaluates store potential) and quantitative (store sales). When offline staff evaluate stores, each person has different standards, resulting in inconsistent data. If store sales are used as the evaluation standard, it is difficult to reasonably select products for stores with potential but poor performance, resulting in the loss of a large number of potential channel users. Summary of the invention

[0004] Based on this, the present application provides a method, device, equipment and storage medium for selecting products for store shelves to solve the technical problem that it is difficult to reasonably select products for shelves in stores with potential but poor performance, resulting in the loss of a large number of potential channel users.

[0005] In a first aspect, a method for selecting products for sale in a store is provided, the method comprising:

[0006] Obtain historical sales information of pre-selected stores and geo-tag information of all buildings within a preset range, wherein the geo-tag information includes geo-tag category, number of geo-tags corresponding to pre-selected stores, and average distance between each geo-tag and the pre-selected stores;

[0007] According to the functional types of the geo-tags in the geo-tag categories, the geo-tag categories are merged to obtain the valid tag categories;

[0008] According to the number of geographic tags corresponding to the pre-selected stores and the average distance between each geographic tag and the pre-selected stores, determine the standardized tag distance value between each pre-selected store and the geographic tags in the valid tag category and the standardized tag quantity value of the geographic tags in the valid tag category of the pre-selected stores;

[0009] Determine the classification information of the pre-selected stores according to the standardized distance value and the standardized label quantity value;

[0010] The target stores are identified for listing based on the classification information and historical sales information of the pre-selected stores. The target stores are those whose sales have not reached the preset standards.

[0011] According to an achievable method in an embodiment of the present application, the geo-tag category includes a medium-granularity tag category, and the geo-tag category is merged according to the functional type of the geo-tag in the geo-tag category to obtain a valid tag category, including:

[0012] The geographic tags with the same functional type are merged into the same medium-grained tag category to obtain the valid tag category.

[0013] According to an achievable method in an embodiment of the present application, the valid tag category includes a fine-grained valid tag category and a medium-grained valid tag category; the number of geographic tags corresponding to the pre-selected stores and the average distance between each geographic tag and the pre-selected stores, and determining the standardized tag distance value between each pre-selected store and the geographic tags in the valid tag category include:

[0014] Determine the effective distance value between the fine-grained effective label category and the pre-selected store according to the average distance between each fine-grained effective label category and the pre-selected store;

[0015] Determine the final distance value between each medium-granularity effective label category and the pre-selected store according to the effective distance value, the number of medium-granularity effective label categories, and the number of fine-granularity effective label categories contained in each medium-granularity effective label category;

[0016] According to the final distance value between each medium-granularity effective label category and the pre-selected store, the medium-granularity effective label category is decentralized to obtain the standardized label distance value between each pre-selected store and the geographic label in the effective label category.

[0017] According to an achievable method in an embodiment of the present application, the valid tag category includes a medium-granularity valid tag category; and according to the number of geographic tags corresponding to the pre-selected store, determining the standardized tag quantity value of the geographic tags in the valid tag category of the pre-selected store includes:

[0018] Determine the horizontal proportion of each medium-granularity valid label category in the pre-selected stores in the number of geographic labels corresponding to the pre-selected stores;

[0019] Determine the label weight of each medium-granularity valid label category according to the total number of pre-selected stores and the number of pre-selected stores containing medium-granularity valid label categories;

[0020] According to the horizontal proportion and label weight of each medium-granularity valid label category, the standardized label quantity value of the geographic label in the valid label category of the pre-selected store is determined.

[0021] According to an achievable method in an embodiment of the present application, the classification information of the pre-selected store is determined according to the standardized distance value and the standardized tag quantity value, including:

[0022] Filter the medium-granularity valid label categories whose standardized distance values ​​and standardized label quantity values ​​are both greater than the preset values ​​as the final valid label categories;

[0023] According to the final valid label category, hierarchical clustering is performed on the pre-selected stores to obtain the initial grouping results of the pre-selected stores;

[0024] The initial grouping results of the pre-selected stores are used as the initial center points of the k-means clustering to cluster the pre-selected stores and obtain the classification information of the pre-selected stores.

[0025] According to an achievable method in an embodiment of the present application, the classification information includes store grouping, and the historical sales information includes sales commodities, sales prices, and sales volumes; and determining the listed commodities of the target store according to the classification information and historical sales information of the pre-selected store includes:

[0026] Verify whether the classification information of the pre-selected stores is reasonable based on historical sales information;

[0027] If the classification information of the pre-selected stores is reasonable, the merchandise to be put on the shelves of the target stores will be determined based on the sales products and sales prices of the top stores in the store grouping. The top stores are the pre-selected stores whose sales exceed the preset sales.

[0028] According to an achievable method in the embodiment of the present application, verifying whether the classification information of the pre-selected store is reasonable based on the historical sales information includes:

[0029] Based on the sales price and sales volume of all categories of goods sold in the pre-selected stores, determine the historical top goods whose sales share exceeds the first preset sales share;

[0030] According to the sales price and sales volume of all categories of goods sold in the pre-selected stores in the store grouping, determine the pre-selected top goods whose sales ratio exceeds the second preset sales ratio;

[0031] Based on the types of historical top products and pre-selected top products, verify whether the classification information of the pre-selected stores is reasonable.

[0032] In a second aspect, a device for selecting goods for sale in a store is provided, the device comprising:

[0033] An acquisition module is used to acquire historical sales information of pre-selected stores and geo-tag information of all buildings within a preset range, wherein the geo-tag information includes geo-tag category, the number of geo-tags corresponding to the pre-selected stores, and the average distance between each geo-tag and the pre-selected store;

[0034] A merging module, used for merging geographic tag categories according to functional types of geographic tags in the geographic tag categories to obtain valid tag categories;

[0035] A determination module, used to determine a standardized tag distance value between each pre-selected store and a geotag in a valid tag category and a standardized tag quantity value of the geotag in a valid tag category of the pre-selected store according to the number of geotags corresponding to the pre-selected stores and the average distance between each geotag and the pre-selected store;

[0036] The determination module is further used to determine the classification information of the pre-selected store according to the standardized distance value and the standardized label quantity value;

[0037] The determination module is also used to determine the target stores' shelf merchandise based on the pre-selected stores' classification information and historical sales information. The target stores are stores whose sales have not reached the preset standards.

[0038] In a third aspect, a computer device is provided, comprising:

[0039] at least one processor; and

[0040] a memory communicatively connected to the at least one processor; wherein,

[0041] The memory stores computer instructions that can be executed by the at least one processor, and the computer instructions are executed by the at least one processor to enable the at least one processor to perform the method involved in the first aspect above.

[0042] In a fourth aspect, a computer-readable storage medium is provided, on which computer instructions are stored, characterized in that the computer instructions are used to enable a computer to execute the method involved in the first aspect above.

[0043] According to the technical content provided in the embodiments of the present application, historical sales information of pre-selected stores and geographic tag information of all buildings within a preset range are obtained, wherein the geographic tag information includes geographic tag categories, the number of geographic tags corresponding to pre-selected stores, and the average distance between each geographic tag and the pre-selected store; the geographic tag categories are merged according to the functional types of the geotags in the geotag categories to obtain valid tag categories; the standardized tag distance value between each pre-selected store and the geographic tags in the valid tag category and the standardized tag quantity value of the geotags in the valid tag category of the pre-selected store are determined according to the number of geographic tags corresponding to the pre-selected stores and the average distance between each geographic tag and the pre-selected store; the classification information of the pre-selected stores is determined according to the standardized distance value and the standardized tag quantity value; the listed products of the target stores are determined according to the classification information and historical sales information of the pre-selected stores, so as to optimize the listed products of stores with poor sales performance, recommend products suitable for store sales, and increase store sales. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 This is an application environment diagram of a method for selecting products for store shelves in one embodiment;

[0045] Figure 2 A schematic diagram of a flow chart of a method for selecting products for sale in a store in one embodiment;

[0046] Figure 3 This is a structural block diagram of a device for selecting goods for sale in a store in one embodiment;

[0047] Figure 4 FIG. 4 is a schematic structural diagram of a computer device in one embodiment. DETAILED DESCRIPTION

[0048] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0049] At present, the retail industry is mainly divided into online and offline. As the last link of the entire sales channel, offline terminals have the characteristics of thousands of stores with thousands of faces. The randomness of consumption increases the difficulty of drawing store consumer portraits. How to help a large number of convenience stores choose suitable products for their sales on limited shelf space, reduce trial and error costs, and increase sales has become the focus of store operations. In order to control operating costs, we cannot make personalized strategies for all stores. Therefore, we hope to divide stores into different categories for easy operation and management. By building store portraits for store groups, we can adjust marketing strategies in a targeted manner to improve the effectiveness of strategies and sales.

[0050] At present, when the market segments stores, it is usually based on qualitative (staff evaluates store potential) and quantitative (store sales). When offline staff evaluate stores, each person has different standards, resulting in inconsistent data. When store sales are used as the evaluation standard, sales are closely related to the number of product SKUs on the shelves and the overall sales of the store. As a brand that hopes to help channel dealers increase sales, it is unable to find out how to improve the marketing plans of stores with potential but poor performance, and ultimately loses a large number of potential channel users.

[0051] In order to solve the problems of the prior art, the embodiments of the present application provide a method, device, equipment and computer storage medium for selecting products for sale in a store. For ease of understanding, the system to which the present application is applicable is first described. The method for selecting products for sale in a store provided by the present application can be applied to Figure 1In the system architecture shown. Among them, the terminal 110 communicates with the server 120 through the network. The terminal 110 obtains the historical sales information of the pre-selected store and the geo-tag information of all buildings within the preset range, wherein the geo-tag information includes the geo-tag category, the number of geo-tags corresponding to the pre-selected store, and the average distance between each geo-tag and the pre-selected store. According to the functional type of the geo-tag category, the geo-tag category is merged to obtain the effective tag category. According to the number of geo-tags corresponding to the pre-selected store and the average distance between each geo-tag and the pre-selected store, the standardized tag distance value between each pre-selected store and the geo-tag in the effective tag category and the standardized tag quantity value of the geo-tag in the effective tag category of the pre-selected store are determined. According to the standardized distance value and the standardized tag quantity value, the classification information of the pre-selected store is determined, and according to the classification information and historical sales information of the pre-selected store, the on-shelf goods of the target store are determined. Among them, the terminal 110 can be, but not limited to, various personal computers, laptops, smart phones and tablet computers, and the server 120 can be implemented by an independent server or a server cluster composed of multiple servers.

[0052] Figure 2 A flowchart of a method for selecting products for store shelves provided in an embodiment of the present application, the method can be performed as follows: Figure 1 The terminal 110 in the system shown in the figure executes. Figure 2 As shown, the method may include the following steps:

[0053] S210, obtaining historical sales information of pre-selected stores and geographic tag information of all buildings within a preset range.

[0054] Among them, the pre-selected stores are stores pre-selected for optimizing marketing plans, covering various sales situations, including stores with good sales performance, average sales performance, and poor sales performance.

[0055] Historical sales information is statistical data on sales records and transaction information of pre-selected stores over a period of time in the past, which may include sales quantity, sales revenue, sales price, sales channel, sales time and customer information of listed products.

[0056] The preset range is a range with a certain distance as the radius and the pre-selected store as the center. The geo-tag information includes the geo-tag category, the number of geo-tags corresponding to the pre-selected stores, and the average distance between each geo-tag and the pre-selected stores. The geo-tag category can include fine-grained tag category, medium-grained tag category, and coarse-grained tag category.

[0057] The geo-tag information is obtained through the map application. The location of the pre-selected store is found according to the coordinates of the pre-selected store, and the geo-tags of all buildings within a certain distance radius from the pre-selected store are obtained. The map application can be Amap, and the radius of the preset range can be 150 meters. Taking Amap as an example, Amap adopts a three-level tag system. The fine-grained tags contain 869 detailed geographic annotations, the medium-grained tags are summarized into 264, and the coarse-grained tags are reduced to 24. The objective data obtained through the map application reduces the subjective bias of manual annotation and provides more accurate environmental mapping.

[0058] S220 , merging geographic tag categories according to functional types of the geographic tags in the geographic tag categories to obtain valid tag categories.

[0059] Since the 869 fine-grained label categories are too detailed and contain labels that are irrelevant to the sales of the pre-selected stores, such as natural place names, red lights, etc., and there are a large number of missing labels for the pre-selected stores, resulting in data sparsity problems.

[0060] The coarse-grained layer has only 24 types of labels, which will merge a large number of distinct labels into one category. For example, shopping malls, home appliance and electronics stores, supermarkets, furniture and building materials markets, comprehensive markets, and stationery stores are merged into shopping services. In fact, these labels have obvious differences in function and scale. For example, the store area and the number of shoppers that can be accommodated in a stationery store are much smaller than those in a supermarket. Therefore, the number of labels under the 24 types of labels cannot accurately reflect the characteristics of the labels near the store, and thus it is impossible to accurately classify the stores.

[0061] Medium-granularity tags can separate each geographic tag more accurately, that is, each geographic tag can more accurately describe the geographical features near the pre-selected store, and also provide competitor information, such as shopping malls, comprehensive markets, convenience stores, supermarkets, etc.

[0062] In order to efficiently integrate information, based on the grouping logic of Amap and combined with common sense, a medium-granularity level is used, and some geographic tags with the same functional types are merged. For example, the car repair tags of brands such as GAC Honda Repair and Dongfeng Honda Repair are merged into Car Repair. The final merged valid tag categories are 94 categories.

[0063] By rationalizing geographic annotations, information overload caused by dense annotations is avoided, thereby improving the practicality of the data.

[0064] S230, determining a standardized tag distance value between each pre-selected store and a geographical tag in a valid tag category and a standardized tag quantity value of geographical tags in a valid tag category of the pre-selected store based on the number of geographical tags corresponding to the pre-selected stores and the average distance between each geographical tag and the pre-selected stores.

[0065] The number of each geotag comes from the map application user's self-upload and official certification. When the geotags that change quickly change, such as Chinese restaurants, the map application will not be updated in time. Therefore, the number of such geotags will be greater than the actual number, and there will be hundreds of Chinese restaurants in one geographical location. In addition, for large transportation hubs, such as airports and train stations, the map application only has 1 and 0 annotations, but this type of geotag can bring more traffic than the tags that change quickly. Logically speaking, the higher the number of annotations, the faster the location changes, and the more people look for the location, the greater the traffic it can bring. However, the number of geotags that change quickly (Chinese restaurants) must be much higher than the geotags that change slowly (large transportation hubs). However, the traffic that these two types of geotags can attract and the size of the impact on the store traffic are opposite to the size of the annotation number. If the number of geotags can roughly reflect the number of traffic, that is, the degree of impact on the pre-selected stores, it is necessary to process the distribution of the number of tags as a whole, so as to filter out the tags that have a greater impact on the pre-selected stores.

[0066] In addition, considering the actual situation of customer flow, the distance between the geotag and the pre-selected store becomes a key factor affecting the probability of customer visit: the farther the geotag is, the lower the possibility of customers attracted by the building to visit the pre-selected store. For example, students in school are more likely to visit the store opposite the school rather than the store 100 meters away from the school.

[0067] Since clustering models are used to classify pre-selected stores, and such models mainly classify by calculating the distance between eigenvalues, the size of the eigenvalues ​​themselves will have a greater impact on the clustering results. Therefore, the number of geographic tags and the distance between geographic tags and stores are comprehensively considered to select geographic tags that are outstanding in both quantity and distance, that is, high-quality tags that can represent regional vitality and are close to stores, for store clustering.

[0068] The number of geotags corresponding to the pre-selected stores and the average distance between each geotag and the pre-selected stores determine the standardized tag distance value between each pre-selected store and the geotags in the valid tag category. Based on the number of geotags corresponding to the pre-selected stores, the standardized tag quantity value of the geotags in the valid tag category of the pre-selected stores is determined.

[0069] S240, determining classification information of the pre-selected store according to the standardized distance value and the standardized label quantity value.

[0070] After obtaining the standardized distance value and the standardized tag quantity value, due to the different distribution and units of the geotag quantity and geotag distance value, if the geotag quantity and geotag distance value are directly added or multiplied, the distribution of the values ​​will change, and the weight of the geotag value increases the difficulty of clustering. Because the density and number of buildings in different regions are different, for example, the distances of various geotags in first-tier cities are generally closer and the number is larger, while the number and distance distribution of various tags in second- and third-tier cities are more dispersed. Therefore, when consumers have consumption needs, consumers in first-tier cities have more choices and a higher probability of patronizing stores that are closer, while consumers in second- and third-tier cities can only choose to walk a certain distance to meet their needs, so even if the distance is far, there is still a probability of visiting. This distribution exists objectively, and the value is affected by the region and cannot reflect its true weight. Therefore, it is necessary to simplify the weight of the feature value and only screen the geotags that have a greater impact on the store in terms of the number of geotags and the distance of geotags.

[0071] According to the filtered geographic tags, the pre-selected stores are clustered by combining k-means and hierarchical clustering to obtain the classification information of the pre-selected stores, including store grouping, store sales products, sales prices and other information.

[0072] Since k-means clustering is easily affected by the selection of initial points, its stability is limited, and hierarchical clustering is stable but often difficult to use when facing highly complex data sets. Therefore, we combine the characteristics of the two models and use the grouping results of hierarchical clustering as the starting point of kmeans to cluster, ensuring the stability and reproducibility of the clustering results.

[0073] S250, determining the products to be put on the shelves of the target store based on the classification information and historical sales information of the pre-selected store.

[0074] Among them, the target stores are stores whose sales do not meet the preset standards. The preset standards are sales baselines set according to the store managers' expectations for store profits.

[0075] The rationality of store grouping is verified through the classification information and historical sales information of the pre-selected stores. If the grouping of the pre-selected stores is relatively reasonable, the products on the shelves and sales prices of the top stores in each group are analyzed, and the products are sorted by sales volume from the dimensions of price, packaging, specifications, and variety. The top-selling products are recommended to the weaker performing stores, and the company will give them appropriate preference when investing in marketing activities such as display, so as to increase the store owners' willingness to put products on the shelves and the success rate.

[0076] It can be seen that the embodiment of the present application obtains the historical sales information of the pre-selected store and the geo-tag information of all buildings within a preset range, wherein the geo-tag information includes the geo-tag category, the number of geo-tags corresponding to the pre-selected store, and the average distance between each geo-tag and the pre-selected store, and merges the geo-tag categories according to the functional type of the geo-tags in the geo-tag category to obtain the effective tag category, and determines the standardized tag distance value between each pre-selected store and the geo-tag in the effective tag category and the standardized tag quantity value of the geo-tag in the effective tag category of the pre-selected store according to the number of geo-tags corresponding to the pre-selected store and the average distance between each geo-tag and the pre-selected store, and determines the classification information of the pre-selected store according to the standardized distance value and the standardized tag quantity value, and determines the listed products of the target store according to the classification information and historical sales information of the pre-selected store, so as to optimize the listed products of stores with poor sales performance, recommend products suitable for store sales, and increase store sales.

[0077] As an achievable approach, the geotag categories are merged according to the functional types of the geotags in the geotag categories to obtain valid tag categories, including:

[0078] The geographic tags with the same functional type are merged into the same medium-grained tag category to obtain the valid tag category.

[0079] The functional type of a geographic tag is a classification of the functions or duties of the building corresponding to the geographic tag. For example, the geographic tags are China Construction Bank ATM, Industrial Bank ATM, China Postal Savings Bank ATM, etc., and their functions are all used for self-service deposits and withdrawals. This type of geographic tags can be merged into a medium-granularity tag category such as ATM. The geographic tags in the original medium-granularity tag category are merged according to the functional type to obtain the final valid tag category. Among them, the valid tag category includes fine-grained valid tag categories and medium-granularity valid tag categories, and each medium-granularity valid tag category includes at least one fine-grained valid tag category.

[0080] As an achievable approach, the number of geographic tags corresponding to the pre-selected stores and the average distance between each geographic tag and the pre-selected stores are used to determine the standardized tag distance value between each pre-selected store and the geographic tags in the valid tag category, including:

[0081] Determine the effective distance value between the fine-grained effective label category and the pre-selected store according to the average distance between each fine-grained effective label category and the pre-selected store;

[0082] Determine the final distance value between each medium-granularity effective label category and the pre-selected store according to the effective distance value, the number of medium-granularity effective label categories, and the number of fine-granularity effective label categories contained in each medium-granularity effective label category;

[0083] According to the final distance value between each medium-granularity effective label category and the pre-selected store, the medium-granularity effective label category is decentralized to obtain the standardized label distance value between each pre-selected store and the geographic label in the effective label category.

[0084] Since the number of fine-grained labels corresponding to each medium-grained label is inconsistent, when summarizing the distance values ​​of 630 categories of fine-grained labels into 94 categories of medium-grained labels, it is impossible to directly add them according to the correspondence between medium-grained labels and fine-grained labels. For example, there will be a difference in quantity between a medium-grained label containing 10 fine-grained labels and a medium-grained label containing 5 fine-grained labels, so the number of categories corresponding to medium-grained labels and fine-grained labels also needs to be considered. Therefore, we calculate the distance value of 94 categories of medium-grained labels for each store according to the formula below to reduce the impact of the number of label correspondences.

[0085] According to the average distance between each fine-grained effective label and the pre-selected store, the effective distance value between the fine-grained effective label and the pre-selected store is determined based on formula (1):

[0086]

[0087] Among them, distance nk represents the effective distance between the kth fine-grained effective label category and the pre-selected store n, k is 630, distance ni represents the average distance between the i-th fine-grained valid label category and the pre-selected store n, and J represents the total number of fine-grained valid label categories.

[0088] According to the effective distance value, the number of medium-grained effective label categories, and the number of fine-grained effective label categories contained in each medium-grained effective label category, the final distance value between each medium-grained effective label category and the pre-selected store is determined based on formula (2):

[0089]

[0090] Among them, final nm Indicates the final distance value between the mth medium-granularity effective label category and the pre-selected store n, subcount nm Indicates the number of fine-grained valid label categories included in the m-th medium-grained valid label category of pre-selected store n, totalcount n It indicates the number of medium-granularity valid label categories contained in the pre-selected store n, and s indicates the number of valid distance values.

[0091] The final distance value between the medium-grained effective label category and the pre-selected stores is a relatively reasonable distance data, but the impact of the change in the distance value of each label on the store still needs to be considered. For example, if the distance between the gym and all stores is similar, then the impact of this distance on the store is similar. When the store is classified, the impact of this feature value on the store is small. Therefore, by vertically and horizontally centering the medium-grained effective label category to remove the dimension of the value, the real impact of the feature value on the store is presented with a value. Vertical centering is to remove the dimension of the label itself from the perspective of the label, and unify the final distance value of the medium-grained effective label category into a distribution with an average value of 0, that is, the labels with smaller overall value changes become smaller, and the labels with larger overall value changes become larger; horizontal centering is to screen out the medium-grained effective labels with outstanding numbers from the perspective of the store. The final distance value of this part of the medium-grained effective label category is the standardized label distance value between each pre-selected store and the geographic label in the effective label category.

[0092] As an achievable approach, according to the number of geographic tags corresponding to the pre-selected stores, a standardized tag quantity value of the geographic tags in the valid tag category of the pre-selected stores is determined, including:

[0093] Determine the horizontal proportion of each medium-granularity valid label category in the pre-selected stores in the number of geographic labels corresponding to the pre-selected stores;

[0094] Determine the label weight of each medium-granularity valid label category according to the total number of pre-selected stores and the number of pre-selected stores containing medium-granularity valid label categories;

[0095] According to the horizontal proportion and label weight of each medium-granularity valid label category, the standardized label quantity value of the geographic label in the valid label category of the pre-selected store is determined.

[0096] The 630 fine-grained valid label categories are aggregated into 94 medium-grained valid label categories according to the correspondence between medium-grained labels and fine-grained labels. The label values ​​need to be recalculated and the distance between label values ​​adjusted.

[0097] Due to the different characteristics of the tags themselves, the possibility and frequency of being labeled are also different, that is, the overall value of the geographical tag (Chinese restaurant) that changes faster will be much greater than the geographical tag (airport) that changes slower. Therefore, it is necessary to combine the horizontal proportion of the geographical tag with the probability of the geographical tag itself to reassign the number of tags. For example, for a geographical tag (airport) that changes less and appears less frequently, its number of tags is 1, which means that the horizontal proportion value is a positive number close to 0, and the tag weight value is large, so the multiplied standardized number of tags will be greater than the original 1.

[0098] The horizontal proportion is the proportion of a medium-granularity valid tag category of a pre-selected store in the number of all geographic tags of the pre-selected store.

[0099] The horizontal ratio can be calculated based on formula (3):

[0100]

[0101] Among them, a nm Indicates the horizontal proportion of medium-granularity effective label categories, loc nm It represents the number of the m-th medium-granularity valid label categories of the pre-selected store n, and m represents the number of medium-granularity valid label categories.

[0102] According to the total number of pre-selected stores and the number of pre-selected stores containing medium-granularity valid label categories, the label weight of each medium-granularity valid label category is calculated based on formula (4):

[0103]

[0104] Among them, b m represents the label weight of the m-th medium-granularity effective label category, N store Represents the total number of pre-selected stores, N mstore Represents the number of pre-selected stores n that contain the m-th medium-granularity valid label category.

[0105] Multiply the horizontal proportion of each medium-granularity effective tag category by the tag weight to get a new value for the number of geographic tags. In order to unify the dimension of the tag value, filter out the higher tag values ​​near the pre-selected stores, and then perform vertical and horizontal centralization, and retain tag values ​​greater than 0. Vertical centralization is to remove the dimension from the tag dimension (unify all tags into a distribution with a mean of 0); horizontal centralization is to use numerical values ​​to reflect the tags with prominent numbers from the perspective of a single store, and obtain the standardized tag quantity value of the geographic tags in the effective tag category.

[0106] As an achievable method, the classification information of the pre-selected stores is determined according to the standardized distance value and the standardized label quantity value, including:

[0107] Filter the medium-granularity valid label categories whose standardized distance values ​​and standardized label quantity values ​​are both greater than the preset values ​​as the final valid label categories;

[0108] According to the final valid label category, hierarchical clustering is performed on the pre-selected stores to obtain the initial grouping results of the pre-selected stores;

[0109] The initial grouping results of the pre-selected stores are used as the initial center points of the k-means clustering to cluster the pre-selected stores and obtain the classification information of the pre-selected stores.

[0110] Since the distribution and units of the standardized distance value and the standardized label quantity value are different, if the standardized distance value and the standardized label quantity value are directly added or multiplied, the distribution of the values ​​will change, and the weight of the geographic label value increases the difficulty of clustering. Because the density and number of buildings in different regions are different, for example, the distances of various labels in first-tier cities are generally closer and the number is larger, while the number and distance distribution of various labels in second- and third-tier cities are more dispersed. Therefore, when consumers have consumption needs, consumers in first-tier cities have more choices and are more likely to visit stores that are closer, while consumers in second- and third-tier cities can only choose to walk a certain distance to meet their needs, so even if the distance is far, there is still a probability of visiting.

[0111] This distribution exists objectively. Therefore, after the distance and number of geographic tags are vertically and horizontally centralized, the numerical values ​​are affected by the region and cannot reflect their true weights. By simplifying the weights of the feature values, only the geographic tags that have a greater impact on the pre-selected stores in terms of distance and quantity can be screened and assigned the same value.

[0112] The preset value is set to 0, and the medium-grained valid label categories with standardized distance values ​​and standardized label quantity values ​​both greater than 0 are selected as the final valid label categories and assigned a value of 1, and the others are 0.

[0113] After obtaining the final valid label categories and their values, we cluster the pre-selected stores. The stability of k-means clustering is limited by the fact that it is easily affected by the selection of the initial point; while hierarchical clustering is stable, it is often difficult to exert its effectiveness when facing highly complex data sets. Therefore, we combine the characteristics of the two models and use the grouping results of hierarchical clustering as the starting point of kmeans to cluster, ensuring the stability and reproducibility of the clustering results.

[0114] According to the final valid label category, use hierarchical clustering to preliminarily group the pre-selected stores, and determine the termination conditions of hierarchical clustering, such as reaching a predetermined number of clusters or stopping clustering according to a certain clustering quality indicator. Extract the results of hierarchical clustering to obtain preliminary cluster divisions, that is, the initial grouping results of the pre-selected stores.

[0115] Preset the sampling ratio of k-means clustering, the number of clusters, the initial value for measuring k-means results (the sum of the overall distance squares accounts for the sum of the overall distance squares), the number of calculations, the distance calculation method, etc. The result of hierarchical clustering is used as the initial center point of k-means clustering. Specifically, for each cluster obtained by hierarchical clustering, calculate its center point, that is, the mean of all points in the cluster, and then use these center points as the initial point of k-means clustering. Execute the k-means clustering algorithm, and when the sum of the overall distance squares output by the model accounts for a smaller proportion of the total distance squares than the initial value, update it to the new initial value until the center point no longer changes significantly. Based on the results of k-means clustering, obtain the classification information of the pre-selected stores.

[0116] Among them, the calculation methods of the standardized distance values ​​of the geographic tags in the final valid tag category include Euclidean distance, Manhattan distance, Pearson correlation coefficient distance, etc. The calculation method of the standardized distance value can be selected as needed, and the Euclidean distance can be used to define the similarity between clusters.

[0117] In addition, the optimal number of clusters can be determined by the elbow method and the silhouette coefficient. The elbow method determines the optimal number of clusters by plotting the sum of squared errors of each cluster as the number of clusters increases. The silhouette coefficient takes into account the compactness within the cluster and the separation between clusters, and calculates the average silhouette coefficient of each cluster. And according to the clustering results and the difficulty of actual operation, the number of categories is selected. Due to limited resources, if there are too many store categories, it is not convenient for brands and offline operators to manage. Therefore, different numbers of clusters are tried in combination with the elbow method, silhouette coefficient and operational needs. Because the number of stores is too large, in order to improve the efficiency of calculation, modeling is carried out by sampling (70%). At the same time, in order to optimize the model results, we use the ratio between the group spacing and the overall distance of the clustering results to measure the clustering results of each sampling, and find the optimal solution for the clustering results in the loop.

[0118] As a feasible method, the target store's listing products are determined based on the pre-selected store's classification information and historical sales information, including:

[0119] Verify whether the classification information of the pre-selected stores is reasonable based on historical sales information;

[0120] If the classification information of the pre-selected stores is reasonable, the merchandise to be put on the shelves of the target stores will be determined based on the sales products, sales prices and sales volumes of the top stores in the store grouping. The top stores are the pre-selected stores whose sales exceed the preset sales.

[0121] The classification information includes store groups, and the historical sales information includes sales commodities, sales prices, and sales volumes. Based on the sales commodities of all categories of the pre-selected stores, their sales prices, and sales volumes, check whether there are obvious differences in the sales of commodities of each category in different groups of the pre-selected stores according to the store groups of the pre-selected stores.

[0122] Specifically, based on the sales price and sales volume of all categories of goods sold in the pre-selected stores, determine the historical top goods whose sales ratio exceeds the first preset sales ratio;

[0123] According to the sales price and sales volume of all categories of goods sold in the pre-selected stores in the store grouping, determine the pre-selected top goods whose sales ratio exceeds the second preset sales ratio;

[0124] Based on the types of historical top products and pre-selected top products, verify whether the classification information of the pre-selected stores is reasonable.

[0125] The first preset sales ratio is the ratio of the sales of the product in the sales of all categories when it is confirmed as a hot-selling product in the historical sales process of a pre-selected store. The sales price and sales volume of each product in all categories of the pre-selected store are multiplied to obtain the sales volume, and the sales ratio of the product is calculated. If the sales ratio exceeds the first preset sales ratio, the product is considered to be a hot-selling product and is used as the historical top product.

[0126] The second preset sales ratio is the ratio of the sales of all categories of goods in all pre-selected stores when the sales product is confirmed as a hot-selling product in the historical sales process of the pre-selected stores in the store group. The sales price and sales volume of each sales product in all categories of sales products in the pre-selected stores are multiplied to obtain the sales volume, the sales of the same sales products in the pre-selected stores in the same group are accumulated, and the ratio of the sales product in the sales of all pre-selected stores in the group is calculated. If the ratio exceeds the second preset sales ratio, the sales product is considered to be a hot-selling product and is used as the pre-selected top product.

[0127] Compare whether the types of historical head products and pre-selected head products are the same or similar. If the types of historical head products and pre-selected head products are the same or similar, it means that there is almost no difference in the sales of each category in the pre-selected stores in this group, indicating that the classification information of the pre-selected stores is reasonable. If the types of historical head products and pre-selected head products are not the same or similar, it means that there is a significant difference in the sales of each category in the pre-selected stores in this group, indicating that the classification information of the pre-selected stores is unreasonable. For example, according to past sales history, in stores near primary and secondary schools, the sales of snacks account for a high proportion, in stores near hospitals, infant-related products are more popular, and in stores near leisure and entertainment, consumers are more inclined to buy functional beverage products. If the pre-selected head products of the stores in the store group are also these, it means that the classification information of the pre-selected stores is reasonable. If they are not very relevant, it means that the classification information of the pre-selected stores does not meet the actual sales needs and is unreasonable.

[0128] If the classification information of the pre-selected stores is reasonable, the products to be put on the shelves of the target stores will be determined based on the sales products, sales prices and sales volumes of the top stores in the store grouping.

[0129] The head store is a pre-selected store with sales exceeding the preset sales. Sales exceeding the preset sales indicate that the store is a pre-selected store with relatively high sales. The preset sales can be set according to profit needs. The sales commodities and sales prices of the head store shelves in each category are analyzed, and the prices, packaging, specifications, and varieties of the commodities are comprehensively considered and sorted by sales volume. The commodities with high sales volume are regarded as the head hot-selling commodities and recommended to the stores with weaker sales performance as the shelves. At the same time, when the company invests in the display and other marketing activities expenses, it will be appropriately tilted to increase the store owner's willingness and success rate to put the products on the shelves.

[0130] This embodiment deletes and merges fine-grained label categories, cleans the number of geographic labels and the distance dimension between labels and stores as a whole, selects labels that can highlight the geographical characteristics of stores, and then combines k-means and hierarchical clustering to mine geographic labels near stores to ensure the stability of clustering results, optimize the company's marketing activity policies based on the results, and provide a basis for optimizing channel dealers' product shelves. An important feature that distinguishes it from other data methods is that objective geographic labels can effectively reflect the environment around the store. Buildings in the surrounding environment will attract consumers of different sizes and needs. For example, consumers near schools and consumers near hospitals will have different needs for infant products. Stores are clustered according to the objective environmental labels around the stores, and to ensure the rationality of the clustering results, the clustering results are verified with sales data of all categories. The clustering results are used to set up product listing plans for stores and guide store product signing.

[0131] It should be understood that although Figure 2The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in the application, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0132] Figure 3 A schematic diagram of a device for selecting goods for sale in a store provided in an embodiment of the present application. The device can be arranged at Figure 1 The terminal 110 in the system shown is used to perform the following operations: Figure 2 The method flow shown in Figure 3 As shown, the device may include: an acquisition module 310, a merging module 320 and a determination module 330, and may further include: a verification module. The main functions of each component module are as follows:

[0133] An acquisition module 310 is used to acquire historical sales information of a pre-selected store and geo-tag information of all buildings within a preset range, wherein the geo-tag information includes geo-tag categories, the number of geo-tags corresponding to the pre-selected stores, and the average distance between each geo-tag and the pre-selected store;

[0134] A merging module 320, configured to merge geographic tag categories according to functional types of the geographic tags in the geographic tag categories to obtain valid tag categories;

[0135] A determination module 330 is used to determine a standardized tag distance value between each pre-selected store and a geotag in a valid tag category and a standardized tag quantity value of the geotag in a valid tag category of the pre-selected store based on the number of geotags corresponding to the pre-selected store and the average distance between each geotag and the pre-selected store;

[0136] The determination module 330 is further used to determine the classification information of the pre-selected store according to the standardized distance value and the standardized label quantity value;

[0137] The determination module 330 is also used to determine the target store's shelf merchandise based on the classification information and historical sales information of the pre-selected stores, where the target store is a store whose sales volume does not reach a preset standard.

[0138] As an implementable approach, the geo-tag categories include medium-granularity tag categories. According to the functional types of the geo-tags in the geo-tag categories, the geo-tag categories are merged to obtain effective tag categories, including:

[0139] The geographic tags with the same functional type are merged into the same medium-grained tag category to obtain the valid tag category.

[0140] As an implementable manner, the effective label category includes a fine-grained effective label category and a medium-grained effective label category; the determination module 330 is specifically used to: determine the effective distance value between the fine-grained effective label category and the pre-selected store according to the average distance between each fine-grained effective label category and the pre-selected store; determine the final distance value between each medium-grained effective label category and the pre-selected store according to the effective distance value, the number of medium-grained effective label categories, and the number of fine-grained effective label categories contained in each medium-grained effective label category; according to the final distance value between each medium-grained effective label category and the pre-selected store, the medium-grained effective label category is decentralized to obtain the standardized label distance value between each pre-selected store and the geographic label in the effective label category.

[0141] As an implementable method, the effective tag category includes a medium-granularity effective tag category; the determination module 330 is specifically used to: determine the horizontal proportion of each medium-granularity effective tag category in the number of geographic tags corresponding to the pre-selected stores; determine the tag weight of each medium-granularity effective tag category based on the total number of pre-selected stores and the number of pre-selected stores containing the medium-granularity effective tag category; determine the standardized tag quantity value of the geographic tags in the effective tag category of the pre-selected stores based on the horizontal proportion and tag weight of each medium-granularity effective tag category.

[0142] As an achievable method, the determination module 330 is specifically used to: screen the medium-granularity effective label categories whose standardized distance values ​​and standardized label quantity values ​​are both greater than preset values ​​as the final effective label categories; perform hierarchical clustering on the pre-selected stores according to the final effective label categories to obtain the initial grouping results of the pre-selected stores; cluster the pre-selected stores using the initial grouping results of the pre-selected stores as the initial center points of the k-means clustering to obtain the classification information of the pre-selected stores.

[0143] As a feasible method, the classification information includes store grouping, and the determination module 330 is specifically used to: verify whether the classification information of the pre-selected store is reasonable based on the historical sales information; if the classification information of the pre-selected store is reasonable, determine the listed goods of the target store based on the sales products and sales prices of the head store in the store grouping, and the head store is the pre-selected store whose sales exceed the preset sales.

[0144] As a feasible method, the device also includes a verification module, which is used to: determine the historical head commodity whose sales share exceeds the first preset sales share based on the sales price and sales volume of all categories of commodities sold in the pre-selected store; determine the pre-selected head commodity whose sales share exceeds the second preset sales share based on the sales price and sales volume of all categories of commodities sold in the pre-selected store in the store grouping; and verify whether the classification information of the pre-selected store is reasonable based on the types of historical head commodities and pre-selected head commodities.

[0145] The same or similar parts between the above embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0146] It should be noted that the embodiments of the present application may involve the use of user data. In actual applications, user-specific personal data can be used in the scheme described in this article within the scope permitted by applicable laws and regulations, provided that the applicable laws and regulations of the country are met (for example, the user's explicit consent, effective notification to the user, and the user's explicit authorization, etc.).

[0147] According to an embodiment of the present application, the present application also provides a computer device and a computer-readable storage medium.

[0148] like Figure 4 , is a block diagram of a computer device according to an embodiment of the present application. The computer device is intended to represent various forms of digital computers or mobile devices. The digital computer may include a desktop computer, a portable computer, a workbench, a personal digital assistant, a server, a mainframe computer, and other suitable computers. The mobile device may include a tablet computer, a smart phone, a wearable device, etc.

[0149] like Figure 4 As shown, the computer device 400 includes a computing unit 401, a ROM 402, a RAM 403, a bus 404, and an input / output (I / O) interface 405. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other through the bus 404. The input / output (I / O) interface 405 is also connected to the bus 404.

[0150] The computing unit 401 can perform various processes in the method embodiment of the present application according to the computer instructions stored in the read-only memory (ROM) 402 or the computer instructions loaded from the storage unit 408 to the random access memory (RAM) 403. The computing unit 401 can be various general and / or special processing components with processing and computing capabilities. The computing unit 401 may include, but is not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. In some embodiments, the method provided in the embodiment of the present application can be implemented as a computer software program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 408.

[0151] RAM 403 may also store various programs and data required for the operation of computer device 400. Part or all of the computer program may be loaded and / or installed on computer device 400 via ROM 402 and / or communication unit 409.

[0152] The input unit 406, output unit 407, storage unit 408 and communication unit 409 in the computer device 400 can be connected to the I / O interface 405. The input unit 406 can be, for example, a keyboard, a mouse, a touch screen, a microphone, etc. The output unit 407 can be, for example, a display, a speaker, an indicator light, etc. The computer device 400 can exchange information, data, etc. with other devices through the communication unit 409.

[0153] It should be noted that the device may also include other components necessary for normal operation, or may only include components necessary for implementing the solution of the present application, rather than all the components shown in the figure.

[0154] Various implementations of the systems and techniques described herein can be realized in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof.

[0155] The computer instructions for implementing the method of the present application can be written in any combination of one or more programming languages. These computer instructions can be provided to the computing unit 401, so that when the computer instructions are executed by the computing unit 401 such as a processor, the steps involved in the method embodiment of the present application are executed.

[0156] The computer-readable storage medium provided in the present application may be a tangible medium that may contain or store computer instructions for executing the steps involved in the method embodiments of the present application. The computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, and other forms of storage media.

[0157] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A method for selecting goods for sale in a store, characterized in that: The method comprises: Obtaining historical sales information of pre-selected stores and geo-tag information of all buildings within a preset range, wherein the geo-tag information includes geo-tag category, the number of geo-tags corresponding to the pre-selected stores, and the average distance between each geo-tag and the pre-selected stores; According to the functional types of the geotags in the geotag categories, the geotag categories are merged to obtain a valid tag category; According to the number of geographical tags corresponding to the pre-selected stores and the average distance between each geographical tag and the pre-selected stores, determine the standardized tag distance value between each pre-selected store and the geographical tags in the valid tag category and the standardized tag quantity value of the geographical tags in the valid tag category of the pre-selected stores; Determining classification information of the pre-selected store according to the standardized distance value and the standardized label quantity value; The commodities to be put on the shelves of the target stores are determined according to the classification information of the pre-selected stores and the historical sales information, and the target stores are stores whose sales have not reached the preset standard.

2. The method according to claim 1, characterized in that The geotag categories include medium-granularity tag categories, and the geotag categories are merged according to the functional types of the geotags in the geotag categories to obtain valid tag categories, including: The geographic tags with the same functional type are merged into the same medium-grained tag category to obtain the valid tag category.

3. The method according to claim 1, characterized in that The effective tag categories include fine-grained effective tag categories and medium-grained effective tag categories; the number of geographic tags corresponding to the pre-selected stores and the average distance between each geographic tag and the pre-selected stores, and determining the standardized tag distance value between each pre-selected store and the geographic tags in the effective tag categories, including: Determine the effective distance value between the fine-grained effective label category and the pre-selected store according to the average distance between each of the fine-grained effective label categories and the pre-selected store; Determine a final distance value between each of the medium-granularity valid label categories and the pre-selected store according to the effective distance value, the number of the medium-granularity valid label categories, and the number of fine-granularity valid label categories contained in each of the medium-granularity valid label categories; According to the final distance value between each of the medium-granularity valid tag categories and the pre-selected stores, the medium-granularity valid tag categories are decentralized to obtain a standardized tag distance value between each pre-selected store and the geographic tag in the valid tag category.

4. The method according to claim 1, characterized in that: The valid tag category includes a medium-granularity valid tag category; and determining the standardized tag quantity value of the geographic tags in the valid tag category of the pre-selected store according to the number of geographic tags corresponding to the pre-selected store, including: Determine the horizontal proportion of each of the medium-granularity valid tag categories in the pre-selected stores in the number of geographic tags corresponding to the pre-selected stores; Determining a label weight of each medium-granularity valid label category according to the total number of the pre-selected stores and the number of the pre-selected stores containing the medium-granularity valid label category; According to the horizontal proportion and label weight of each of the medium-granularity valid label categories, the standardized label quantity value of the geographic label in the valid label category of the pre-selected store is determined.

5. The method according to claim 1, characterized in that The determining the classification information of the pre-selected store according to the standardized distance value and the standardized label quantity value includes: Selecting a medium-granularity valid label category whose standardized distance value and standardized label quantity value are both greater than a preset value as a final valid label category; According to the final valid label category, hierarchical clustering is performed on the pre-selected stores to obtain an initial grouping result of the pre-selected stores; The initial grouping results of the pre-selected stores are used as initial center points of k-means clustering to cluster the pre-selected stores, so as to obtain classification information of the pre-selected stores.

6. The method according to claim 1, characterized in that The classification information includes store grouping, and the historical sales information includes sales commodities, sales prices and sales volumes; The step of determining the listed commodities of the target store according to the classification information of the pre-selected store and the historical sales information includes: Verify whether the classification information of the pre-selected store is reasonable based on the historical sales information; If the classification information of the pre-selected stores is reasonable, the merchandise on the shelves of the target store is determined based on the sales products and sales prices of the head store in the store grouping, and the head store is a pre-selected store whose sales exceed the preset sales.

7. The method according to claim 6, characterized in that Verifying whether the classification information of the pre-selected store is reasonable based on the historical sales information includes: Determine, based on the sales price and sales volume of all categories of goods sold in the pre-selected store, the historical top goods whose sales ratio exceeds the first preset sales ratio; Determine, based on the sales price and sales volume of all categories of goods sold in the pre-selected stores in the store group, the pre-selected top goods whose sales ratio exceeds the second preset sales ratio; Based on the types of the historical head products and the pre-selected head products, verify whether the classification information of the pre-selected stores is reasonable.

8. A device for selecting goods for sale in a store, characterized in that: The device comprises: An acquisition module, used to acquire historical sales information of a pre-selected store and geographic tag information of all buildings within a preset range, wherein the geographic tag information includes the geographic tag category, the number of geographic tags corresponding to the pre-selected store, and the average distance between each geographic tag and the pre-selected store; A merging module, used for merging the geotag categories according to the functional types of the geotags in the geotag categories to obtain a valid tag category; A determination module, configured to determine, according to the number of geographic tags corresponding to the pre-selected stores and the average distance between each geographic tag and the pre-selected stores, a standardized tag distance value between each pre-selected store and the geographic tags in the valid tag category and a standardized tag quantity value of the geographic tags in the valid tag category of the pre-selected stores; The determination module is further used to determine the classification information of the pre-selected store according to the standardized distance value and the standardized label quantity value; The determination module is also used to determine the listed goods of the target store based on the classification information of the pre-selected store and the historical sales information, and the target store is a store whose sales volume does not reach the preset standard.

9. A computer device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores computer instructions that can be executed by the at least one processor, and the computer instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer instructions stored thereon, characterized in that: The computer instructions are used to cause a computer to execute the method according to any one of claims 1 to 7.

Images