Merchant matching method, device, equipment and storage medium

Abstract

This application discloses a merchant matching method, apparatus, device, and storage medium, belonging to the field of data processing. The method includes: retrieving a set of points of interest (POIs) from a geographic information system based on at least a portion of the merchant attribute information of the source merchants obtained from a source merchant data source; determining the spatial attribute similarity and non-spatial attribute similarity between the source merchants and the POIs based on the merchant attribute information and the POI attribute information; determining the similarity between the source merchants and the POIs based on the spatial attribute similarity, non-spatial attribute similarity, preset spatial attribute weighting coefficients, and preset non-spatial attribute weighting coefficients; and determining that the source merchants and the POIs are the same merchant if the similarity meets preset matching conditions. The embodiments of this application can improve the accuracy of determining whether merchants corresponding to data from different data sources are the same merchant.

Description

Technical Field

[0001] This application belongs to the field of data processing, and in particular relates to a merchant matching method, apparatus, device and storage medium. Background Technology

[0002] With the development of electronic payment technology, more and more merchants can use it for payments, necessitating the management of merchant data for better electronic payment processing. However, with the popularization of big data technology, data sources are increasingly diverse, often yielding data for the same merchant from multiple data sources. Due to differences in format and description among these data sources, it's possible to mistakenly identify the same merchant as two or more. To reduce the probability of misidentifying the same merchant, merchant matching can be performed to determine if two or more merchants belong to the same entity. However, currently, the accuracy of merchant matching is low, making it difficult to accurately determine whether merchants corresponding to data from different data sources belong to the same entity. Summary of the Invention

[0003] This application provides a merchant matching method, apparatus, device, and storage medium, which can improve the accuracy of determining whether merchants corresponding to data from different data sources belong to the same merchant.

[0004] In a first aspect, embodiments of this application provide a merchant matching method, comprising: retrieving a set of points of interest (POIs) in a geographic information system based on at least a portion of the merchant attribute information of a source merchant obtained from a source merchant data source, the set of POIs including at least one POI; determining the spatial attribute similarity and non-spatial attribute similarity between the source merchant and the POIs based on the merchant attribute information and the POIs' POI attribute information, the non-spatial attribute similarity being obtained based on at least two non-spatial sub-attribute similarities; determining the similarity between the source merchant and the POIs based on the spatial attribute similarity, non-spatial attribute similarity, a preset spatial attribute weight coefficient, and a preset non-spatial attribute weight coefficient; and determining that the source merchant and the POIs are the same merchant if the similarity meets preset matching conditions.

[0005] Secondly, embodiments of this application provide a merchant matching device, comprising: a retrieval module, configured to retrieve a set of points of interest (POIs) in a geographic information system based on at least a portion of the merchant attribute information of the source merchant obtained from a source merchant data source, the set of POIs including at least one POI; a first similarity calculation module, configured to determine the spatial attribute similarity and non-spatial attribute similarity between the source merchant and the POIs based on the merchant attribute information and the POIs's POI attribute information, the non-spatial attribute similarity being obtained based on at least two non-spatial sub-attribute similarities; a second similarity calculation module, configured to determine the similarity between the source merchant and the POIs based on the spatial attribute similarity, non-spatial attribute similarity, a preset spatial attribute weight coefficient, and a preset non-spatial attribute weight coefficient; and a determination module, configured to determine that the source merchant and the POIs are the same merchant if the similarity meets preset matching conditions.

[0006] Thirdly, embodiments of this application provide an electronic device, including: a processor and a memory storing computer program instructions; the processor executes the computer program instructions to implement the merchant matching method of the first aspect.

[0007] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the merchant matching method of the first aspect.

[0008] This application provides a merchant matching method, apparatus, device, and storage medium. Based on merchant attribute information of source merchants obtained from a source merchant data source, a set of points of interest (POIs) is retrieved from a geographic information system. Based on the merchant attribute information and the POI attribute information of the POIs in the POI set, the spatial attribute similarity and non-spatial attribute similarity between the source merchant and the POIs are determined. The similarity between the source merchant and the POIs is comprehensively obtained based on the spatial attribute similarity, non-spatial attribute similarity, spatial attribute weight coefficient, and non-spatial attribute weight coefficient. Finally, based on the similarity and preset matching conditions, it is determined whether the source merchant and the POIs are the same merchant. Spatial attribute similarity and non-spatial attribute similarity describe the degree of similarity between source merchants and points of interest from different dimensions. Non-spatial attribute similarity is obtained based on at least two non-spatial sub-attribute similarities. Different non-spatial sub-attribute similarities have different dimensions and describe the degree of non-spatial similarity between source merchants and points of interest from different dimensions. The similarity between source merchants and points of interest is obtained by the similarity between source merchants and points of interest in multiple dimensions, which is more accurate. Using this similarity to determine whether source merchants and points of interest are the same merchant can improve the accuracy of judging whether merchants corresponding to data from different data sources are the same merchant. Attached Figure Description

[0009] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 A flowchart illustrating a merchant matching method provided in an embodiment of this application;

[0011] Figure 2 A flowchart illustrating a merchant matching method provided in another embodiment of this application;

[0012] Figure 3 This is a schematic diagram illustrating an example of a shared category relationship tree in an embodiment of this application;

[0013] Figure 4 This is a schematic diagram illustrating an example of the hierarchical relationship of category information in the source merchant data source in this application embodiment;

[0014] Figure 5 This is a schematic diagram illustrating another example of the hierarchical relationship of category information in the source merchant data source in this application embodiment;

[0015] Figure 6 A flowchart illustrating a merchant matching method provided in yet another embodiment of this application;

[0016] Figure 7 This is a schematic diagram illustrating another example of the hierarchical relationship of category information in the source merchant data source in this application embodiment;

[0017] Figure 8 This is a schematic diagram illustrating another example of the hierarchical relationship of category information in a geographic information system as described in this application.

[0018] Figure 9 This is a schematic diagram illustrating another example of a shared category relationship tree in the embodiments of this application;

[0019] Figure 10 This is a schematic diagram of the structure of a merchant matching device provided in an embodiment of this application;

[0020] Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0021] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0022] With the development of electronic payment technology, more and more merchants can use it for payments, necessitating the management of merchant data for better electronic payment processing. However, with the popularization of big data technology, data sources are increasingly diverse, often yielding data for the same merchant from multiple data sources. Due to differences in format and description among these data sources, it's possible to mistakenly identify the same merchant as two or more. To reduce the probability of misidentifying the same merchant, merchant matching can be performed to determine if two or more merchants belong to the same entity. However, currently, the accuracy of merchant matching is low, making it difficult to accurately determine whether merchants corresponding to data from different data sources belong to the same entity.

[0023] This application provides a merchant matching method, apparatus, device, and storage medium, capable of obtaining merchant attribute information of source merchants from a source merchant data source, and retrieving a set of points of interest from a geographic information system based on at least a portion of this information. Based on the merchant attribute information of the source merchants and the point of interest attribute information, the spatial and non-spatial attributes of the source merchants and the points of interest in the set are fully explored, with non-spatial attributes further subdivided into multiple dimensions. The similarity between the source merchants and the points of interest can be calculated based on the spatial and non-spatial attributes; this similarity is derived from information across multiple dimensions, resulting in higher accuracy. Using this similarity to determine whether a source merchant and a point of interest belong to the same merchant can also improve the accuracy of determining whether merchants corresponding to data from different data sources belong to the same merchant. The merchant matching method, apparatus, device, and storage medium provided in this application can be applied to scenarios such as merchant deduplication and merchant authenticity verification, and are not limited thereto.

[0024] It should be noted that the acquisition, storage, use, and processing of information and data in this application embodiment are all authorized by the user or relevant organization and comply with the relevant provisions of national laws and regulations. The merchant matching method, apparatus, equipment, and storage medium provided in this application will be described below.

[0025] The first aspect of this application provides a merchant matching method that can be applied to a merchant matching device or equipment, that is, the merchant matching method can be executed by a merchant matching device or equipment, and is not limited thereto. Figure 1 A flowchart of a merchant matching method provided in an embodiment of this application is shown below. Figure 1 As shown, the merchant matching method may include steps S101 to S104.

[0026] In step S101, a set of points of interest is retrieved from the geographic information system based on at least a portion of the merchant attribute information of the source merchants obtained from the source merchant data source.

[0027] The source merchant data source is the merchant's own data source; the data in the source merchant data source comes directly from the source merchant itself. Merchant attribute information can be used to characterize the merchant's attributes, and the source merchant's attribute information can be used to characterize the source merchant's attributes. Merchant attribute information may include multiple aspects of the source merchant's information. In some examples, merchant attribute information may include, but is not limited to, basic merchant information, merchant address information, and merchant category information. Basic merchant information may include, but is not limited to, one or more of the following: merchant name, unified social credit code, business registration name, legal representative's name, contact person's name, and contact telephone number. Merchant address information can be used to characterize the merchant's address; for example, merchant address information may include address elements such as the merchant's province, city, district, and house number. Merchant category information can be used to define the merchant's category; for example, merchant categories may include supermarkets, convenience stores, restaurants, Western restaurants, Chinese restaurants, shopping services, farmers' markets, fruit shops, building materials, catering services, and gourmet food, etc. The classification method for merchants is not limited here.

[0028] A Geographic Information System (GIS) is an information system that stores map information, provides electronic maps, and enables the querying of geographic information. Based on GIS, point-of-interest (POI) retrieval functions such as keyword search and surrounding area search can be implemented. In some examples, at least some of the merchant's attribute information can be used as search criteria to perform a retrieval within the GIS. For instance, a keyword search can be used, with the city in the merchant's name and address information as search parameters, to retrieve POIs similar to the merchant's name within a fixed area. A surrounding area search can also be used, with the latitude and longitude of the merchant's address information, the search range, and the merchant's name as search parameters, to retrieve POIs similar to the merchant's name within a certain distance around a fixed latitude and longitude. A combination of keyword search and surrounding area search can also be used. The retrieved POIs form a set of POIs, meaning the set of POIs includes at least one POI. In some examples, information associated with merchant attribute information can be obtained based on merchant attribute information. This associated information can then be used as search criteria to retrieve a set of points of interest (POIs) in a geographic information system (GIS). Alternatively, at least a portion of the merchant attribute information and associated information can be used as search criteria to retrieve the same set of POIs. The specific merchant attribute information and associated information used as search criteria can be determined based on the scenario and the emphasis of the requirements, and are not limited here. Specifically, geographic search parameters can be obtained based on merchant attribute information. These parameters can then be used to perform several rounds of keyword searches and / or surrounding area searches to obtain the set of POIs. Geographic search parameters include information that can be used for geographic searches; for example, they may include, but are not limited to, merchant names and address information. When at least two rounds of keyword searches and / or surrounding area searches are performed based on merchant attribute information, the union of the POIs obtained from these two rounds can be defined as the POI set. For example, based on merchant attribute information, n rounds of retrieval were performed, and the subsets of interest points obtained from the n rounds of retrieval were P1, P2, ..., P... n Then the set of interest points P = P1∪P2∪……∪P n .

[0029] In some examples, when obtaining merchant attribute information, verification can also be performed on this information. Verification may include consistency verification, accuracy verification, and completeness verification, and is not limited here. The verification results can determine whether to continue with subsequent steps, or adjust the weighting coefficients involved in subsequent similarity calculations. For example, if merchant attribute information includes a unified social credit code, registered business name, and legal representative's name, the unified social credit code can be used to query the enterprise map to see if the registered business name and legal representative's name corresponding to that unified social credit code in the enterprise map match the registered business name and legal representative's name in the merchant attribute information. If they match, the consistency verification is considered passed; if they do not match, the consistency verification is considered failed. As another example, if merchant attribute information includes merchant address information, it can be determined whether the merchant address information includes address elements such as province, city, district, and house number for completeness verification; or it can be verified based on whether the accuracy of the merchant address information meets a preset verification accuracy standard for accuracy verification.

[0030] In step S102, based on merchant attribute information and point of interest attribute information, the spatial attribute similarity and non-spatial attribute similarity between the source merchant and the point of interest are determined.

[0031] Point-of-interest (POI) attribute information can be used to characterize the attributes of a POI. POI attribute information may include multiple aspects of the POI; in some examples, it may include, but is not limited to, basic POI information, POI address information, and POI category information. Basic POI information may include, but is not limited to, one or more of the following: POI name, Unified Social Credit Code, registered business name, legal representative's name, contact person's name, and contact telephone number. POI address information can be used to characterize the POI's address; for example, it may include the merchant's province, city, district, and house number. POI category information can be used to classify the merchant; for example, the merchant category may include supermarkets, convenience stores, restaurants, Western restaurants, Chinese restaurants, shopping services, farmers' markets, fruit shops, building materials stores, catering services, and gourmet food. The classification method for POIs is not limited here.

[0032] Spatial attribute similarity characterizes the similarity between a source merchant and a point of interest (POI) in the spatial dimension, referring to the proximity of a pair of entities in geographic space. Non-spatial attribute similarity can include the similarity between the source merchant and POI in non-spatial dimensions, and can be obtained based on at least two non-spatial sub-attribute similarities. In some embodiments, non-spatial sub-attribute similarities may include text similarity, address similarity, category similarity, etc., and are not limited thereto. Text similarity characterizes the similarity between the merchant attribute information of the source merchant and the POI attribute information in the text dimension, used for information comparison at the pure text level. Address similarity characterizes the similarity between the source merchant and POI in the address dimension. Category similarity characterizes the similarity between the source merchant and POI in the category dimension.

[0033] Spatial attribute similarity and non-spatial attribute similarity can describe the degree of similarity between source merchants and points of interest from different dimensions. Non-spatial attribute similarity is obtained based on at least two non-spatial sub-attribute similarities. Different non-spatial sub-attribute similarities have different dimensions, and different non-spatial sub-attribute similarities also describe the degree of non-spatial similarity between source merchants and points of interest from different dimensions.

[0034] In step S103, the similarity between the source merchant and the point of interest is determined based on the spatial attribute similarity, non-spatial attribute similarity, preset spatial attribute weight coefficient, and preset non-spatial attribute weight coefficient.

[0035] Spatial attribute weight coefficients are the weight coefficients corresponding to spatial attribute similarity, and can be set according to specific scenarios, needs, experience, etc., and are not limited here. Non-spatial attribute weight coefficients are the weight coefficients corresponding to non-spatial attribute similarity, and can be set according to specific scenarios, needs, experience, etc., and are not limited here.

[0036] The similarity between the source merchant and the point of interest is a comprehensive similarity calculated by combining spatial attribute similarity and non-spatial attribute similarity. The higher the similarity, the greater the likelihood that the source merchant and the point of interest are the same merchant. The similarity between the source merchant and the point of interest can be calculated by weighting spatial attribute similarity and non-spatial attribute similarity. For example, the similarity between the source merchant and the point of interest can be calculated according to the following formula (1):

[0037] (1)

[0038] in, The similarity between source merchants and points of interest; Spatial attribute similarity; These are the spatial attribute weighting coefficients; This refers to non-spatial attribute similarity. These are non-spatial attribute weighting coefficients.

[0039] In step S104, if the similarity meets the preset matching conditions, it is determined that the source merchant and the point of interest are the same merchant.

[0040] The preset matching conditions include those used to determine whether a source merchant and a point of interest are the same merchant based on similarity. If the similarity does not meet the preset matching conditions, the source merchant and the point of interest are determined to be different users. For example, preset matching conditions include a similarity greater than a preset similarity threshold. The preset similarity threshold can be set according to specific scenarios and needs, and is not limited here. If the similarity is greater than the preset similarity threshold, the source merchant and the point of interest can be determined to be the same merchant; if the similarity is less than or equal to the preset similarity threshold, the source merchant and the point of interest can be determined to be different merchants.

[0041] For each point of interest in the set of points of interest, steps S102 to S104 above can be executed to obtain the similarity between the source merchant and each point of interest, thereby determining whether there are any points of interest in the set of points of interest that are the same merchant as the source merchant.

[0042] In some examples, before step S103 above, corresponding similarity judgment thresholds can be set for spatial attribute similarity, non-spatial attribute similarity, and non-spatial sub-attribute similarity, respectively, according to requirements. If the spatial attribute similarity, non-spatial attribute similarity, and non-spatial sub-attribute similarity are all greater than the corresponding similarity judgment thresholds, steps S103 and S104 can be continued. If the similarity of any one of the spatial attribute similarity, non-spatial attribute similarity, and non-spatial sub-attribute similarity is less than or equal to the corresponding similarity judgment threshold, it can be directly determined that the source merchant and the point of interest are different merchants.

[0043] In this embodiment, based on the merchant attribute information of the source merchant obtained from the source merchant data source, a set of points of interest is retrieved in the geographic information system. Based on the merchant attribute information and the point of interest attribute information of the points of interest in the set, the spatial attribute similarity and non-spatial attribute similarity between the source merchant and the points of interest are determined. According to the spatial attribute similarity, non-spatial attribute similarity, spatial attribute weight coefficient, and non-spatial attribute weight coefficient, the similarity between the source merchant and the points of interest is comprehensively obtained. Then, based on the similarity and preset matching conditions, it is determined whether the source merchant and the points of interest are the same merchant. Spatial attribute similarity and non-spatial attribute similarity describe the degree of similarity between source merchants and points of interest from different dimensions. Non-spatial attribute similarity is obtained based on at least two non-spatial sub-attribute similarities. Different non-spatial sub-attribute similarities have different dimensions and describe the degree of non-spatial similarity between source merchants and points of interest from different dimensions. The similarity between source merchants and points of interest is obtained by the similarity between source merchants and points of interest in multiple dimensions, which is more accurate. Using this similarity to determine whether source merchants and points of interest are the same merchant can improve the accuracy of judging whether merchants corresponding to data from different data sources are the same merchant.

[0044] In some embodiments, there may be situations where the similarity between the source merchant and the points of interest in the set of points of interest does not meet the preset matching criteria. In some examples, when the similarity does not meet the preset matching criteria, the search criteria can be modified to retrieve a new set of points of interest from the geographic information system based on at least a portion of the merchant's attribute information. The modified search criteria have a wider search scope, enabling the retrieval of other points of interest outside the original set of points of interest; that is, the new set of points of interest includes at least one point of interest that does not exist in the original set of points of interest. After obtaining the new set of points of interest, the similarity between the source merchant and the points of interest in the new set of points of interest is determined, and it is determined whether the source merchant and the points of interest in the new set of points of interest are the same merchant. In some examples, when the similarity does not meet the preset matching criteria, the search can be stopped, and it can be determined that there are no points of interest in the geographic information system that match the source merchant.

[0045] In some embodiments, merchant attribute information may include merchant address information, and point of interest attribute information may include point of interest address information. Figure 2 A flowchart illustrating a merchant matching method provided in another embodiment of this application. Figure 2 and Figure 1 The difference is that, Figure 1 Step S102 can be further refined as follows: Figure 2 Steps S1021 to S1023 in the process.

[0046] In step S1021, spatial attribute similarity is determined based on merchant address information and point of interest address information.

[0047] Merchant address information and point-of-interest (POI) address information can be converted into geographic location information. Using this geographic location information, the distance between the source merchant and the POI can be determined. Based on this distance, spatial attribute similarity can be obtained. The smaller the distance between the source merchant and the POI, the greater the spatial attribute similarity. Specifically, merchant address information can be converted into first geographic location information; POI address information can be converted into second geographic location information; based on the first and second geographic location information, a first distance between the source merchant and the POI can be calculated; and spatial attribute similarity can be determined based on the first distance and a preset maximum similarity distance.

[0048] The first geographic location information includes geographic location information converted from merchant address information, and the second geographic location information includes geographic location information converted from point-of-interest (POI) address information. In some examples, the first and second geographic location information can be implemented as coordinates in the same coordinate system. The first distance is the distance between the source merchant and the POI calculated based on the first and second geographic location information. The maximum similarity distance is the maximum deviation between the POI and the source merchant, which may be the same merchant as the source merchant. This can be determined based on specific scenarios, needs, experience, etc., and is not limited here. For example, the maximum similarity distance is 3 kilometers. If the first distance is greater than or equal to the maximum similarity distance, the source merchant and the POI can be considered different merchants, and the spatial attribute similarity can be 0. If the first distance is less than the maximum similarity distance, the spatial attribute similarity can be obtained by the ratio of the difference between the maximum similarity distance and the first distance to the maximum similarity distance.

[0049] For example, the first geographic location information and the second geographic location information can be implemented as longitude and latitude, and the first geographic location information is ( The second geographical location information is ( ), The first distance can be calculated according to the following formula (2), and the spatial attribute similarity can be obtained according to the following formula (3):

[0050] (2)

[0051] (3)

[0052] in, The longitude is the first geographic location information. The latitude is the first geographic location information. The longitude is the second geographic location information; The latitude is the second geographic location information. This is the first distance; The maximum similarity distance; This represents the maximum value of spatial attribute similarity. This represents spatial attribute similarity.

[0053] In step S1022, the non-spatial sub-attribute similarity is determined based on merchant attribute information and point of interest attribute information.

[0054] By comparing different types of merchant attribute information and point-of-interest (POI) attribute information, different non-spatial sub-attribute similarities can be obtained. The smaller the difference between merchant attribute information and POI attribute information in the non-spatial dimension, the greater the non-spatial sub-attribute similarity. Non-spatial sub-attribute similarity can include text similarity, address similarity, category similarity, etc., and is not limited here.

[0055] In some examples, merchant attribute information also includes basic merchant information, point-of-interest attribute information includes basic point-of-interest information, and non-spatial sub-attribute similarity includes text similarity. At least one first keyword for each attribute category can be extracted from the basic merchant information; at least one second keyword for each attribute category can be extracted from the basic point-of-interest information; based on the first and second keywords, the keyword similarity corresponding to each attribute category is obtained; using the keyword similarity corresponding to each attribute category and a preset weight coefficient corresponding to each attribute category, the text similarity between the basic merchant information and the basic point-of-interest information is determined.

[0056] Text processing can be performed on basic merchant information and point-of-interest (POI) information to extract primary and secondary keywords. The primary keyword includes keywords extracted from the basic merchant information, and the secondary keyword includes keywords extracted from the POI information. The method of keyword extraction is not limited here; for example, string matching methods or Natural Language Processing (NLP) methods can be used to determine the primary and secondary keywords for attribute categories in the basic merchant and POI information. Attribute categories can include pre-specified attribute categories, such as brand, type, branch, etc. Based on the primary and secondary keywords, a string similarity algorithm can be used to calculate the keyword similarity between the primary and secondary keywords for each attribute category. Based on the keyword similarity of the primary and secondary keywords for each attribute category, a weighted algorithm can be used to obtain the text similarity. The weight coefficient corresponding to each attribute category can be set according to specific scenarios and needs, and is not limited here. The aforementioned string similarity algorithm may include, but is not limited to, edit distance algorithms and longest consecutive common character algorithms.

[0057] Specifically, if a keyword in the same attribute category does not have a first keyword and / or a second keyword, the keyword similarity to the corresponding attribute category is the minimum similarity value. If the first keyword and the second keyword in the same attribute category have an inclusion relationship, the keyword similarity to the corresponding attribute category is the maximum similarity value. If the first keyword and the second keyword in the same attribute category do not have an inclusion relationship, the keyword similarity to the corresponding attribute category is obtained based on the maximum of the edit distance between the first keyword and the second keyword, and the text character length of the first keyword and the second keyword. The minimum and maximum similarity values ​​can be set according to specific scenarios and requirements, and are not limited here. For example, the minimum similarity value can be 0, and the maximum similarity value can be 1. An inclusion relationship between the first keyword and the second keyword can include the case where the first keyword includes the second keyword and the second keyword includes the first keyword. The case where the first keyword includes the second keyword is the same as the first keyword. Similarly, the case where the second keyword includes the first keyword is the same as the first keyword. In some examples, if the keyword similarity to the corresponding attribute category is the minimum similarity value, the weight coefficient of the corresponding attribute category can be reduced; if the keyword similarity to the corresponding attribute category is the maximum similarity value, the weight coefficient of the corresponding attribute category can be increased.

[0058] For example, the keyword similarity corresponding to an attribute category can be calculated according to the following formula (4), and the text similarity can be calculated according to the following formula (5):

[0059] (4)

[0060] (5)

[0061] in, Let be the keyword similarity corresponding to the i-th attribute category; The first keyword corresponding to the i-th attribute category; The second keyword corresponding to the i-th attribute category; The edit distance between the first keyword and the second keyword corresponding to the i-th attribute category can include the Levenshtein distance. The length of the text characters of the first keyword corresponding to the i-th attribute category; is the length of the text characters of the second keyword corresponding to the i-th attribute category; 0 is the minimum value of keyword similarity, and 1 is the maximum value of keyword similarity. Text similarity; Let be the keyword similarity corresponding to the i-th attribute category; Here is the weight coefficient corresponding to the j-th attribute category; is the weight coefficient corresponding to the i-th attribute category.

[0062] , If at least one of the keywords is missing, that is, if there is no first keyword and / or second keyword under the same attribute category, then the first keyword and second keyword cannot be compared, and therefore the keyword similarity can be 0. , The existence and inclusion relationship indicate that both the first keyword and the second keyword exist under the i-th attribute category, and the objects they point to are the same. Therefore, the keyword similarity can be 1. , If the keyword exists but does not have an inclusion relationship, it means that both the first keyword and the second keyword exist under the i-th attribute category, but the objects they point to may not be the same. Therefore, keyword similarity needs to be calculated based on edit distance and text character length.

[0063] In some examples, non-spatial sub-attribute similarity includes address similarity. Merchant address information can be segmented into words to obtain a first segmentation result; based on the first segmentation result, a first feature vector is constructed using a vector transformation algorithm; point-of-interest address information can be segmented into words to obtain a second segmentation result; based on the second segmentation result, a second feature vector is constructed using a vector transformation algorithm; the similarity between the first and second feature vectors is calculated, and this similarity is determined as the address similarity.

[0064] Address information inherently possesses textual attributes, but it also faces semantic and standardization challenges. Natural language processing methods, such as the HanLP language processing package and the NLPIR-ICTCLAS word segmentation system, can be used to segment merchant address information and point-of-interest (POI) address information into a series of word sequences, yielding a first segmentation result and a second segmentation result. The first segmentation result includes the words derived from the merchant address information, and the second segmentation result includes the words derived from the POI address information. Vector transformation algorithms can convert the segmentation results into feature vectors. The type of vector transformation algorithm is not limited here; for example, based on the first and second segmentation results, the Term Frequency–Inverse Document Frequency (TF-IDF) algorithm can be used to obtain a first feature vector and a second feature vector. The first feature vector is the feature vector obtained from the first segmentation result, and the second feature vector is the feature vector obtained from the second segmentation result. Due to semantic and standardization issues with address information, such as merchant address information including administrative divisions while point-of-interest (POI) address information does not, although they are essentially the same address, the edit distance algorithm results in a low similarity between them, leading to low accuracy in address similarity calculation. To improve the accuracy of address similarity, the cosine similarity algorithm can be used to calculate address similarity. For example, the address similarity between merchant address information and POI address information can be calculated using the following formula (6):

[0065] (6)

[0066] in, Address similarity; This is the first eigenvector; This is the second feature vector; This represents the value of the i-th term in the first eigenvector; Let be the value of the i-th term in the second eigenvector.

[0067] In some examples, merchant attribute information also includes merchant category information, point-of-interest (POI) attribute information also includes POI category information, and non-spatial sub-attribute similarity includes category similarity. A shared category relationship tree between the source merchant data source and the geographic information system can be obtained; within this shared category relationship tree, the first target level where the merchant category information resides and the second target level where the POI category information resides are determined; based on the distance between the first and second target levels and the number of levels in the shared category relationship tree, category similarity is determined.

[0068] For both the source merchant data source and the geographic information system, the number of shared category relationship trees is not limited; there can be one shared category relationship tree or more than two. The shared category relationship tree comprises multiple levels, each level including more than two category information items. The first level can include two category information items and can be considered the root node of the shared category relationship tree. Each category information item in the second level and subsequent levels can be considered a node, and each level can be viewed as a collection of category information. For example, Figure 3 This is a schematic diagram illustrating an example of a shared category relationship tree in an embodiment of this application, such as... Figure 3 As shown, the shared category relationship tree comprises four levels. The first level includes category information a1 and b1; the second level includes category information a2, a3, b2, and b3; the third level includes category information a4, a5, a6, b4, b5, and b6; and the fourth level includes category information b7 and b8. The category information from the source merchant data source includes a1, a2, a3, a4, a5, and a6, while the category information from the geographic information system includes b1, b2, b3, b4, b5, b6, b7, and b8. Both merchant category information and point-of-interest (POI) category information can be found within the levels of the shared category relationship tree. The first target level is the level where the merchant category information resides in the shared category relationship tree. The second target level is the level where the POI category information resides in the shared category relationship tree. The distance between the first and second target levels represents the distance between the merchant category information and the POI category information within the shared category relationship tree. The distance between the first target level and the second target level, and the number of levels in the shared category relationship tree, determine the magnitude of category similarity. When the number of levels in the shared category relationship tree is fixed, the smaller the distance between the first target level and the second target level, the greater the category similarity. When the distance between the first target level and the second target level is fixed, the greater the number of levels in the shared category relationship tree, the greater the category similarity.

[0069] For example, category similarity can be calculated according to the following formulas (7) and (8):

[0070] (7)

[0071] (8)

[0072] in, This refers to the merchant category information at the i-th level in the source merchant data source; This refers to the category information of points of interest located at level j in the geographic information system. Merchant category information Information on points of interest categories The distance between them, that is, the distance between the first target level and the second target level; Merchant category information Hierarchical structure of merchant category information within the shared category relationship tree. and point of interest category information The distance of the shared parent hierarchy; For point of interest category information Hierarchical structure of merchant category information within the shared category relationship tree. and point of interest category information The distance of the shared parent hierarchy; For category similarity; The number of levels in the shared category relationship tree can be determined by the maximum value between the number of levels of category information in the source merchant data source and the number of levels of category information in the geographic information system.

[0073] The shared category relationship tree can be obtained based on the hierarchical relationships of category information in the source merchant data source and the geographic information system. It can be constructed by taking the union of the hierarchical category information in the source merchant data source and the geographic information system. Specifically, the hierarchical relationships of category information in the source merchant data source and the geographic information system can be obtained; semantic recognition can be performed on the category information in the source merchant data source and the geographic information system; if the semantics of the category information in the source merchant data source and the category information in the geographic information system are consistent, the union of the category information in the source merchant data source and the category information in the geographic information system can be constructed according to the hierarchical relationships of the category information to obtain the shared category relationship tree.

[0074] Category information in the source merchant data source has a hierarchical relationship, as does category information in the geographic information system. Based on semantic differences, category information in the source merchant data source and category information in the geographic information system with consistent semantics are grouped into a set. The union of category information at the same level in this set is then obtained to form a common category relationship tree.

[0075] For example, Figure 4 This is a schematic diagram illustrating an example of the hierarchical relationship of category information in the source merchant data source in this application embodiment. Figure 5 This is a schematic diagram illustrating another example of the hierarchical relationship of category information in the source merchant data source in this application embodiment. Figure 4 Category information and Figure 5 The category information in can be constructed as follows Figure 3 The tree showing the common category relationships. Figure 4 The category information a1 in the data is the highest-level category information in the source merchant's data source. Figure 5Category information b1 in the data is the highest-level category information in the geographic information system. Figure 4 The semantics of category information a1 in the middle and Figure 5 The semantics of category information b2 are consistent, and can be obtained according to the hierarchical relationship. Figure 4 Category information and Figure 5 The union of category information. Figure 3 The first level of the class relationship tree includes: Figure 4 The first level of category information and Figure 5 The union of the first-level category information. Figure 3 The second level of the class relationship tree includes Figure 4 The second level of category information and Figure 5 The union of the second-level category information. Figure 3 The third level of the Chinese class relationship tree includes: Figure 4 The third level of category information and Figure 5 The union of the third-level category information, due to Figure 4 The fourth level of category information is not included. Figure 3 The fourth level of the Chinese class relationship tree includes: Figure 5 The third level of category information.

[0076] In the above embodiments, before step S102 or before text processing the merchant basic information and point-of-interest (POI) basic information, text preprocessing can be performed on the merchant attribute information and POI attribute information to obtain merchant attribute information and POI attribute information with consistent character formats. For example, based on a preset dictionary, foreign languages ​​can be uniformly converted to English, foreign languages ​​can be uniformly converted to Chinese, traditional Chinese characters can be uniformly converted to simplified Chinese characters, brand aliases can be uniformly converted to the official brand name, etc., and this is not limited. The merchant attribute information and POI attribute information have a unified character format, which can improve the accuracy of the non-spatial sub-attribute similarity between merchant attribute information and POI attribute information, thereby improving the accuracy of non-spatial attribute similarity.

[0077] In step S1023, the non-spatial attribute similarity is determined based on the non-spatial sub-attribute similarity and the weight coefficient corresponding to the non-spatial sub-attribute similarity.

[0078] Non-spatial attribute similarity can be calculated using a weighted algorithm based on the non-spatial sub-attribute similarity. For example, non-spatial attribute similarity includes text similarity, address similarity, and category similarity. Non-spatial attribute similarity can be calculated using the following formula (9):

[0079] (9)

[0080] in, This refers to non-spatial attribute similarity. Text similarity; These are the weighting coefficients corresponding to text similarity. Address similarity; These are the weighting coefficients corresponding to address similarity; For category similarity; This represents the weighting coefficient corresponding to the category similarity.

[0081] In some embodiments, extended information on points of interest matching the source merchant can be added to the source merchant to enrich the source merchant's profile. Figure 6 A flowchart illustrating a merchant matching method provided in yet another embodiment of this application. Figure 6 and Figure 1 The difference is that, Figure 6 The merchant matching method shown may also include step S105.

[0082] In step S105, if it is determined that the source merchant and the point of interest are the same merchant, the extended information of the point of interest is associated with the source merchant.

[0083] Extended information about points of interest can include data that the source merchant does not have, such as merchant ratings, the business district they belong to, and user reviews. This extended information can be linked to the source merchant and added to the source merchant's data source to enrich the user profile.

[0084] This application fully considers the differences between source merchant data and point of interest data, such as different data structures, incomplete data, information conflicts, naming differences, and inaccurate data. It comprehensively mines spatial and non-spatial attributes and uses a weighted algorithm to calculate the similarity between source merchants and points of interest. The merchant matching method in this application is closer to the business scenario and has higher accuracy.

[0085] To facilitate understanding, the merchant matching process is illustrated below with a specific example, which may include steps c1 to c11.

[0086] In step c1, the merchant attribute information of the source merchant M0 is obtained from the database of the source merchant data source. The merchant attribute information of the source merchant M0 is shown in Table 1 below:

[0087] Table 1

[0088]

[0089] From Table 1, you can obtain merchant attribute information such as merchant name, merchant address, unified social credit code, business registration name, legal representative name, contact person name, and contact number. You can directly use the merchant attribute information, or you can obtain related information based on the merchant attribute information. For example, you can obtain the E1 area where the D1 floor of No. C1, B1 Road is located by associating it with the D1 floor of No. C1, B1 Road.

[0090] In step c2, the merchant attribute information can be verified.

[0091] For example, based on the Unified Social Credit Code XXXXXXXXXXXXXXXXXH in Table 1, a business map can be queried to obtain search results including the Unified Social Credit Code, registered business name, and legal representative's name. The retrieved Unified Social Credit Code, registered business name, and legal representative's name can then be verified for consistency with the Unified Social Credit Code, registered business name, and legal representative's name in the merchant attribute information. Another example is the accuracy verification of merchant addresses in Table 1. The preset accuracy standard is district / county. The merchant addresses in Table 1 are already accurate down to the house number, which is more precise than district / county, and can be considered to have passed the accuracy verification. If the merchant attribute information fails verification, the subsequent merchant matching process can be interrupted or continued; this is not limited to either method.

[0092] In step c3, a set of points of interest is retrieved from the geographic information system based on the merchant attribute information.

[0093] For example, a surrounding search method can be used. The merchant address "B1 Road, C1, Floor D1" in Table 1 above can be resolved into latitude and longitude "x, y". Based on "B1 Road, C1, Floor D1", the associated information "City A1" can be obtained. The search city is confirmed as "City A1", the search area as a "circular area", the search radius as a surrounding "2 kilometers", and the keyword as "X1 Convenience Store (B1 Road Store)". The first round of retrieval from the geographic information system yields two points of interest, denoted as... Points of interest and points of interest The interest point attribute information is shown in Table 2 below:

[0094] Table 2

[0095]

[0096] Alternatively, a keyword search can be used. Confirming the search city as City A1 and the keyword as "XXXXX Convenience Store B1 Road Branch", a second round of retrieval is performed from the geographic information system, yielding one point of interest, denoted as... Points of interest The interest point attribute information is shown in Table 3 below:

[0097] Table 3

[0098]

[0099] Points of interest set .

[0100] In step c4, the source merchants are calculated based on the merchant attribute information and the point of interest attribute information. With points of interest Spatial attribute similarity.

[0101] Assume the maximum similarity distance is 5 kilometers, and the maximum spatial attribute similarity is 1. (The source merchant...) Merchant address information and points of interest The point of interest address information is converted to coordinates in the same coordinate system, such as latitude and longitude coordinates, for the source merchant. With points of interest The latitude and longitude coordinates are shown in Table 4 below:

[0102] Table 4

[0103]

[0104] According to the above formula (2), the source merchants can be calculated. With points of interest The first distances are 0 meters, 3375 meters, and 19067 meters, respectively. According to formula (3) above, the source merchant... With points of interest The spatial attribute similarities were 1, 0.325, and 0, respectively.

[0105] In step c5, the source merchants are calculated based on the merchant basic information and the point of interest basic information. With points of interest Text similarity.

[0106] Text similarity can be calculated based on the merchant name and contact number attributes. The merchant name and contact number can be preprocessed by removing parentheses and characters from the merchant name. Then, using string matching methods or natural language processing word segmentation tools, the first and second keywords for the following four attribute categories can be obtained:

[0107] ={X1, convenience store, B1 road store, XXX-XXXXXXX5};

[0108] ={X2, Convenience Store, B1 Road Store, XXX-XXXXXXX5};

[0109] ={X1, convenience store, B2 Road store, XXX-XXXXXXX9};

[0110] ={X3, self-selected restaurant, B3 Road store, XXX-XXXXXXX3}.

[0111] In the preset dictionary, the three brand keywords X1, X2, and X3 are all mapped to keyword X1. Using the preset dictionary, the second keyword X1, X2, and X3 can all be converted to keyword X1. The source merchant can be calculated using the above formula (4). With points of interest The keyword similarity corresponding to the first attribute category is 1, which can be denoted as S1={1, 1, 1}, source merchant. With points of interest The keyword similarity corresponding to the second attribute category can be denoted as S2={1, 1, 0}, source merchant With points of interest The keyword similarity corresponding to the third attribute category can be denoted as S3={1, 0.5, 0.5}, source merchant. With points of interest The keyword similarity corresponding to the fourth attribute category can be denoted as S4={1, 0, 0}. Assuming the weight coefficients for all four attribute categories are 1, the source merchant can be calculated according to the above formula (5). With points of interest The text similarity is 1, source merchant With points of interest The text similarity is 0.625, source merchant With points of interest The text similarity is 0.375.

[0112] In step c6, the source merchants are calculated based on the merchant address information and the point of interest address information. With points of interest Address similarity.

[0113] Source Merchants With points of interest Taking address similarity calculation as an example, we can first use the Chinese language processing package HanLP to analyze the source merchants. Merchant address information and points of interest The point of interest address information is used for word segmentation, and the first word segmentation result is: ={B1 Road, C1 Number, D1 Floor}, the resulting second word segmentation result is: ={B1 Road, C1 Number}. Based on the term frequency-inverse document frequency algorithm, a feature vector is constructed, and the source merchant... The first feature vector is [1, 1, 1], and the interest point The second feature vector is [1, 1, 0]. According to the above formula (6), the source merchant can be calculated. With points of interest The address similarity is 0.816. Similarly, points of interest... After word segmentation, the second word segmentation result does not contain any word segments identical to those in the first word segmentation result. With points of interest The address similarity is 0, source merchant With points of interest The address similarity is 0.

[0114] In step c7, the source merchants are calculated based on the merchant category information and the point of interest category information, respectively. With points of interest Category similarity.

[0115] Figure 7 This is a schematic diagram illustrating another example of the hierarchical relationship of category information in the source merchant data source in this application embodiment. Figure 8 This is a schematic diagram illustrating another example of the hierarchical relationship of category information in a geographic information system, as described in the embodiments of this application. Figure 9 This is a schematic diagram illustrating another example of a shared category relationship tree in the embodiments of this application. Wherein, Figure 9 It includes two shared category relation trees, namely shared category relation tree U1 and shared category relation tree U2, which are based on Figure 7 The hierarchical relationship of category information in the Zhongyuan merchant data source and Figure 8 The hierarchical relationship of category information in the geographic information system is obtained. Figure 7 "Supermarket" in the middle and Figure 8 If the meaning of "supermarket" in the text is consistent, then it can be... Figure 7 The Chinese text contains terms related to "supermarket" and other related categories. Figure 8 The information on "supermarkets" and other related categories is grouped into one set; Figure 7 The "restaurant" in the text and Figure 8 If the meaning of "catering services" is consistent, then it can be... Figure 7 The Chinese text contains terms related to "restaurants" and other related categories. Figure 8 The category "Food and Beverage Services" and other related categories are grouped into another set. Each set corresponds to a shared category relationship tree. Within each set, the union of category information at the same level is constructed to form the shared category relationship tree corresponding to that set.

[0116] Source Merchants With points of interest The category information is all "convenience store", and the source merchant is... The category information belongs to the second level of the shared category relationship tree U1, and is a point of interest. The category information belongs to the third level of the shared category relationship tree U1, and is the source merchant. Category information and points of interest The parent level for sharing category information is the second level, the source merchant. The distance from the category information to the shared parent level is 0, source merchant The distance from the category information to the shared parent level is 1; Point of Interest The category information is "Supermarket", belonging to the second level of the common category relationship tree U1, and the source merchant. Category information and points of interest The parent level for sharing category information is the second level, points of interest. The distance from the category information to the shared parent level is 0; Point of Interest The category is "food," which belongs to the second level of the shared category relationship tree U2, and is related to the source merchant. The category information does not share a parent hierarchy, therefore points of interest The distance from the category information to the shared parent level is +∞. According to the above formula (7) and (8), the source merchant With points of interest The category similarity is 0.84, source merchant With points of interest The category similarity is 1, source merchant With points of interest The category similarity is 0.

[0117] In step c8, based on the source merchant With points of interest The text similarity, address similarity, category similarity, and the weight coefficients corresponding to text similarity, address similarity, and category similarity are used to calculate the source merchant's value. With points of interest Non-spatial attribute similarity.

[0118] Assuming that the weight coefficients for text similarity, address similarity, and category similarity are all 1, the source merchant can be calculated according to the above formula (9). With points of interest The non-spatial attribute similarity is 0.885, source merchant With points of interest The non-spatial attribute similarity is 0.541, and the source merchant With points of interest The non-spatial attribute similarity is 0.125.

[0119] In step c9, based on the source merchant With points of interest Spatial attribute similarity, non-spatial attribute similarity, spatial attribute weight coefficient, and non-spatial attribute weight coefficient are used to calculate the source merchant. With points of interest The similarity.

[0120] Assuming that both spatial attribute weighting coefficient and non-spatial attribute weighting coefficient are 1, the source merchant can be calculated according to the above formula (1). With points of interest The similarity is 0.9425, source merchant With points of interest The similarity is 0.433, source merchant With points of interest The similarity is 0.0625.

[0121] In step c10, the similarity between the source merchant and the point of interest is compared with a preset similarity threshold, and the source merchant and the point of interest with a similarity greater than the preset similarity threshold are identified as the same merchant.

[0122] If the preset similarity threshold is 0.8, then in the source merchant With points of interest In the similarity score, only the source merchant With points of interest If the similarity is greater than a preset similarity threshold, the source merchant can be identified. With points of interest For the same merchant, source merchant These are real physical merchants that exist on the electronic map.

[0123] If the preset similarity threshold is 0.95, then the source merchant With points of interest The similarity scores do not exceed the preset similarity threshold. You can return to step c3 to change the search criteria, such as expanding the search scope or revising the search keywords, and perform another search in the geographic information system to obtain new points of interest before calculating the source merchants. Similarity to new points of interest.

[0124] In step c11, the points of interest are... Extended information and source merchants Link to source merchants Supplement the information.

[0125] Points of interest Extended information such as ratings, user reviews, average spending, and the business district they belong to is linked to the source merchant. Enrich source merchants Merchant profile information.

[0126] The second aspect of this application provides a merchant matching device. Figure 10 This is a schematic diagram of the structure of a merchant matching device provided in an embodiment of this application, as shown below. Figure 10As shown, the merchant matching device 200 may include a retrieval module 201, a first similarity calculation module 202, a second similarity calculation module 203, and a determination module 204.

[0127] The retrieval module 201 can be used to retrieve a set of points of interest in a geographic information system based on at least a portion of the merchant attribute information of the source merchants obtained from the source merchant data source.

[0128] A set of points of interest includes at least one point of interest.

[0129] The first similarity calculation module 202 can be used to determine the spatial attribute similarity and non-spatial attribute similarity between the source merchant and the point of interest based on the merchant attribute information and the point of interest attribute information.

[0130] Non-spatial attribute similarity is obtained based on at least two non-spatial sub-attribute similarities.

[0131] The second similarity calculation module 203 can be used to determine the similarity between the source merchant and the point of interest based on the spatial attribute similarity, non-spatial attribute similarity, preset spatial attribute weight coefficient, and preset non-spatial attribute weight coefficient.

[0132] The determination module 204 can be used to determine that the source merchant and the point of interest are the same merchant when the similarity meets the preset matching conditions.

[0133] In this embodiment, based on the merchant attribute information of the source merchant obtained from the source merchant data source, a set of points of interest is retrieved in the geographic information system. Based on the merchant attribute information and the point of interest attribute information of the points of interest in the set, the spatial attribute similarity and non-spatial attribute similarity between the source merchant and the points of interest are determined. According to the spatial attribute similarity, non-spatial attribute similarity, spatial attribute weight coefficient, and non-spatial attribute weight coefficient, the similarity between the source merchant and the points of interest is comprehensively obtained. Then, based on the similarity and preset matching conditions, it is determined whether the source merchant and the points of interest are the same merchant. Spatial attribute similarity and non-spatial attribute similarity describe the degree of similarity between source merchants and points of interest from different dimensions. Non-spatial attribute similarity is obtained based on at least two non-spatial sub-attribute similarities. Different non-spatial sub-attribute similarities have different dimensions and describe the degree of non-spatial similarity between source merchants and points of interest from different dimensions. The similarity between source merchants and points of interest is obtained by the similarity between source merchants and points of interest in multiple dimensions, which is more accurate. Using this similarity to determine whether source merchants and points of interest are the same merchant can improve the accuracy of judging whether merchants corresponding to data from different data sources are the same merchant.

[0134] In some embodiments, merchant attribute information includes merchant address information, and point of interest attribute information includes point of interest address information.

[0135] The first similarity calculation module 202 can be used to: determine spatial attribute similarity based on merchant address information and point of interest address information; determine non-spatial sub-attribute similarity based on merchant attribute information and point of interest attribute information; and determine non-spatial attribute similarity based on non-spatial sub-attribute similarity and the weight coefficient corresponding to the non-spatial sub-attribute similarity.

[0136] In some examples, the first similarity calculation module 202 can be used to: convert merchant address information into first geographic location information; convert point of interest address information into second geographic location information; calculate the first distance between the source merchant and the point of interest based on the first geographic location information and the second geographic location information; and determine the spatial attribute similarity based on the first distance and a preset maximum similarity distance.

[0137] In some examples, merchant attribute information also includes basic merchant information, point-of-interest attribute information includes basic point-of-interest information, and non-spatial sub-attribute similarity includes text similarity. The first similarity calculation module 202 can be used to: extract a first keyword for at least one attribute category from the basic merchant information; extract a second keyword for at least one attribute category from the basic point-of-interest information; obtain the keyword similarity corresponding to each attribute category based on the first and second keywords; and determine the text similarity between the basic merchant information and the basic point-of-interest information using the keyword similarity corresponding to each attribute category and the preset weight coefficients corresponding to each attribute category.

[0138] In some examples, the first similarity calculation module 202 can be used to: when there is no first keyword and / or second keyword under the same attribute category, the keyword similarity to the attribute category is the minimum similarity value; when the first keyword and second keyword under the same attribute category have an inclusion relationship, the keyword similarity to the attribute category is the maximum similarity value; when the first keyword and second keyword under the same attribute category do not have an inclusion relationship, the keyword similarity to the attribute category is obtained based on the maximum value of the edit distance between the first keyword and the second keyword, the text character length of the first keyword, and the text character length of the second keyword.

[0139] In some examples, non-spatial sub-attribute similarity includes address similarity. The first similarity calculation module 202 can be used to: segment merchant address information to obtain a first segmentation result; construct a first feature vector based on the first segmentation result using a vector transformation algorithm; segment point of interest address information to obtain a second segmentation result; construct a second feature vector based on the second segmentation result using a vector transformation algorithm; calculate the similarity between the first feature vector and the second feature vector, and determine the similarity between the first feature vector and the second feature vector as the address similarity.

[0140] In some examples, merchant attribute information also includes merchant category information, point-of-interest attribute information also includes point-of-interest category information, and non-spatial sub-attribute similarity includes category similarity. The first similarity calculation module 202 can be used to: obtain a shared category relationship tree between the source merchant data source and the geographic information system, the shared category relationship tree including multiple levels, each level including more than two category information; determine the first target level where the merchant category information is located and the second target level where the point-of-interest category information is located in the shared category relationship tree; and determine the category similarity based on the distance between the first target level and the second target level and the number of levels in the shared category relationship tree.

[0141] In some examples, the merchant matching device 200 may also include a relationship tree construction module. The relationship tree construction module can be used to: obtain the hierarchical relationship of category information in the source merchant data source and the hierarchical relationship of category information in the geographic information system; perform semantic recognition on the category information in the source merchant data source and the category information in the geographic information system; and, if the semantics of the category information in the source merchant data source and the category information in the geographic information system are consistent, construct the union of the category information in the source merchant data source and the category information in the geographic information system according to the hierarchical relationship of the category information, thereby obtaining a common category relationship tree.

[0142] In some embodiments, the merchant matching device 200 may further include a preprocessing module. The preprocessing module may be used to: perform text preprocessing on the merchant attribute information and point-of-interest attribute information to obtain merchant attribute information and point-of-interest attribute information with consistent character format.

[0143] In some embodiments, the retrieval module 201 can also be used to change the retrieval conditions when the similarity does not meet the preset matching conditions, and retrieve a new set of points of interest from the geographic information system based on at least part of the merchant attribute information.

[0144] The new set of points of interest includes at least one point of interest that does not exist in the set of points of interest.

[0145] In some embodiments, the merchant matching device 200 may further include an information association module. The information association module can be used to associate the extended information of the point of interest with the source merchant when it is determined that the source merchant and the point of interest are the same merchant.

[0146] A third aspect of this application also provides an electronic device. Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 11 As shown, the electronic device 300 includes a memory 301, a processor 302, and a computer program stored in the memory 301 and executable on the processor 302.

[0147] In some examples, the processor 302 described above may include a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits that may be configured to implement the embodiments of this application.

[0148] Memory 301 may include mass storage for data or instructions. For example, and not as a limitation, memory 301 may include a hard disk drive (HDD), a floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these.

[0149] Memory 301 may include read-only memory (ROM), random access memory (RAM), disk storage media device, optical storage media device, flash memory device, electrical, optical, or other physical / tangible memory storage device. Therefore, typically, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the merchant matching method according to embodiments of this application.

[0150] The processor 302 runs a computer program corresponding to the executable program code stored in the memory 301 to implement the merchant matching method in the above embodiments.

[0151] In some examples, the electronic device 300 may also include a communication interface 303 and a bus 304. For example, Figure 11 As shown, the memory 301, processor 302, and communication interface 303 are connected through bus 304 and complete communication with each other.

[0152] The communication interface 303 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application. Input devices and / or output devices can also be connected through the communication interface 303.

[0153] Bus 304 includes hardware, software, or both, that couples components of electronic device 300 together. For example, and not as a limitation, bus 304 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-E) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses, or a combination of two or more of these. Where appropriate, bus 304 may include one or more buses. Although specific buses are described and illustrated in the embodiments of this application, this application considers any suitable bus or interconnection.

[0154] A fourth aspect of this application also provides a computer-readable storage medium storing computer program instructions. When these computer program instructions are executed by a processor, they can implement the merchant matching method in the above embodiments and achieve the same technical effect. To avoid repetition, further details are omitted here. The aforementioned computer-readable storage medium may include non-transitory computer-readable storage media, such as read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, etc., and is not limited thereto.

[0155] This application also provides a computer program product. When the instructions in the computer program product are executed by the processor of an electronic device, the electronic device executes the merchant matching method in this application embodiment and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0156] It should be clarified that the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. For the device embodiments, equipment embodiments, and computer-readable storage medium embodiments, the relevant parts can be referred to the description section of the method embodiments. This application is not limited to the specific steps and structures described above and shown in the figures. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application. Furthermore, for the sake of brevity, detailed descriptions of known methods and techniques are omitted here.

[0157] The aspects of this application have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by dedicated hardware performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0158] Those skilled in the art will understand that the above embodiments are exemplary and not restrictive. Different technical features appearing in different embodiments can be combined to achieve beneficial effects. Based on a study of the drawings, specification, and claims, those skilled in the art should be able to understand and implement other variations of the disclosed embodiments. In the claims, the term "comprising" does not exclude other means or steps; the quantifier "a" does not exclude a plurality; the terms "first" and "second" are used to identify names and not to indicate any particular order. No reference numerals in the claims should be construed as limiting the scope of protection. The functionality of multiple parts appearing in the claims can be implemented by a single hardware or software module. The appearance of certain technical features in different dependent claims does not mean that these technical features cannot be combined to achieve beneficial effects.

Claims

1. A merchant matching method, characterized by, The method comprises the following steps: Based on at least part of the merchant attribute information of the source merchant obtained from the source merchant data source, a set of interest points is retrieved in a geographic information system, the set of interest points comprising at least one interest point; Based on the merchant attribute information and the interest point attribute information of the interest point, the spatial attribute similarity and the non-spatial attribute similarity of the source merchant and the interest point are determined, the non-spatial attribute similarity being obtained based on at least two non-spatial sub-attribute similarities; According to the spatial attribute similarity, the non-spatial attribute similarity, a preset spatial attribute weight coefficient and a preset non-spatial attribute weight coefficient of the source merchant and the interest point, the similarity of the source merchant and the interest point is determined; In the case where the similarity meets a preset matching condition, the source merchant and the interest point are determined to be the same merchant; The non-spatial sub-attribute similarity is determined based on the merchant attribute information and the interest point attribute information, the merchant attribute information further comprising merchant category information, the interest point attribute information further comprising interest point category information, and the non-spatial sub-attribute similarity comprising a category similarity, The non-spatial sub-attribute similarity is determined based on the merchant attribute information and the interest point attribute information, comprising: A common category relationship tree of the source merchant data source and the geographic information system is obtained, the common category relationship tree comprising multiple levels, and each level comprising two or more category information; A first target level where the merchant category information is located and a second target level where the interest point category information is located are determined in the common category relationship tree; According to the distance between the first target level and the second target level and the number of levels of the common category relationship tree, the category similarity is determined.

2. The method of claim 1, wherein, The merchant attribute information comprises merchant address information, and the interest point attribute information comprises interest point address information; The spatial attribute similarity and the non-spatial attribute similarity of the source merchant and the interest point are determined based on the merchant attribute information and the interest point attribute information of the interest point, comprising: The spatial attribute similarity is determined based on the merchant address information and the interest point address information; The non-spatial sub-attribute similarity is determined based on the merchant attribute information and the interest point attribute information; The non-spatial attribute similarity is determined according to the non-spatial sub-attribute similarity and a weight coefficient corresponding to the non-spatial sub-attribute similarity.

3. The method of claim 2, wherein, The spatial attribute similarity is determined based on the merchant address information and the interest point address information, comprising: The merchant address information is converted into first geographic position information; The interest point address information is converted into second geographic position information; According to the first geographic position information and the second geographic position information, a first distance between the source merchant and the interest point is calculated; According to the first distance and a preset maximum similarity distance, the spatial attribute similarity is determined.

4. The method of claim 2, wherein, The merchant attribute information further comprises merchant basic information, the interest point attribute information comprises interest point basic information, and the non-spatial sub-attribute similarity comprises a text similarity, The non-spatial sub-attribute similarity is determined based on part of the merchant attribute information and part of the point of interest attribute information, and includes: extracting a first keyword of at least one attribute category from the merchant basic information; extracting a second keyword of at least one attribute category from the point of interest basic information; obtaining a keyword similarity corresponding to each attribute category according to the first keyword and the second keyword; determining the text similarity of the merchant basic information and the point of interest basic information by using the keyword similarity corresponding to each attribute category and a preset weight coefficient corresponding to each attribute category.

5. The method of claim 4, wherein, The keyword similarity corresponding to each attribute category is obtained according to the first keyword and the second keyword, and includes: in the same attribute category, if the first keyword and / or the second keyword is not present, the keyword similarity corresponding to the attribute category is a minimum similarity value; in the same attribute category, if the first keyword and the second keyword have a containing relationship, the keyword similarity corresponding to the attribute category is a maximum similarity value; in the same attribute category, if the first keyword and the second keyword do not have a containing relationship, the keyword similarity corresponding to the attribute category is obtained according to the maximum value of the edit distance of the first keyword and the second keyword, the text character length of the first keyword and the text character length of the second keyword.

6. The method of claim 2, wherein, The non-spatial sub-attribute similarity includes an address similarity, The non-spatial sub-attribute similarity is determined based on part of the merchant attribute information and part of the point of interest attribute information, and includes: performing word segmentation on the merchant address information to obtain a first word segmentation result; constructing a first feature vector by using a vector conversion algorithm according to the first word segmentation result; performing word segmentation on the point of interest address information to obtain a second word segmentation result; constructing a second feature vector by using the vector conversion algorithm according to the second word segmentation result; calculating the similarity of the first feature vector and the second feature vector, and determining the similarity of the first feature vector and the second feature vector as the address similarity.

7. The method of claim 1, wherein, Further comprising: obtaining the hierarchical relationship of the category information in the source merchant data source and the hierarchical relationship of the category information in the geographic information system; performing semantic recognition on the category information in the source merchant data source and the category information in the geographic information system; in the case that the semantics of the category information in the source merchant data source and the semantics of the category information in the geographic information system are consistent, constructing a union set of the category information in the source merchant data source and the category information in the geographic information system according to the hierarchical relationship of the category information, to obtain the common category relationship tree.

8. The method of claim 1, wherein, Before determining the spatial attribute similarity and the non-spatial attribute similarity of the source merchant and the point of interest based on the merchant attribute information and the point of interest attribute information of the point of interest, further comprising: performing text preprocessing on the merchant attribute information and the point of interest attribute information to obtain the merchant attribute information and the point of interest attribute information in the same character format.

9. The method of claim 1, wherein, Further comprising: In a case where the similarity does not satisfy the preset matching condition, the retrieval condition is changed, and a new point of interest set is retrieved from the geographic information system based on at least part of the business attribute information, the new point of interest set including at least one point of interest that does not exist in the point of interest set.

10. The method of claim 1, wherein, Also included are: In a case where the source business and the point of interest are determined to be the same business, the expansion information of the point of interest is associated with the source business.

11. A merchant matching apparatus, comprising: Included are: A retrieval module configured to retrieve a point of interest set from a geographic information system based on at least part of business attribute information of a source business obtained from a source business data source, the point of interest set including at least one point of interest; A first similarity calculation module configured to determine spatial attribute similarity and non-spatial attribute similarity of the source business and the point of interest based on the business attribute information and point of interest attribute information of the point of interest, the non-spatial attribute similarity being based on at least two non-spatial sub-attribute similarities; A second similarity calculation module configured to determine similarity of the source business and the point of interest according to the spatial attribute similarity, the non-spatial attribute similarity, a preset spatial attribute weight coefficient, and a preset non-spatial attribute weight coefficient of the source business and the point of interest; A determination module configured to determine that the source business and the point of interest are the same business in a case where the similarity satisfies a preset matching condition. The non-spatial sub-attribute similarity is determined based on the business attribute information and the point of interest attribute information, the business attribute information further including business category information, and the point of interest attribute information further including point of interest category information, and the non-spatial sub-attribute similarity includes category similarity, The first similarity calculation module is configured to obtain a common category relationship tree of the source business data source and the geographic information system, the common category relationship tree including multiple levels, each level including two or more category information; determine a first target level where the business category information is located and a second target level where the point of interest category information is located in the common category relationship tree; and determine the category similarity according to a distance between the first target level and the second target level and a number of levels of the common category relationship tree.

12. An electronic device, comprising: Included are: A processor and a memory having computer program instructions stored thereon; The processor, when executing the computer program instructions, implements the business matching method of any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that, The computer program instructions are stored on the computer readable storage medium, and when executed by the processor, implement the business matching method of any one of claims 1 to 10.

Images