Verification Method, Device, Equipment and Storage Medium for Real Stores

Abstract

The present application discloses a method, apparatus, device and storage medium for verifying a real store, belonging to the field of data processing. The method includes: obtaining the store information of a first store input; based on the store name and store address of the first store, searching for a first target store name in a first target area in an electronic map, where the first target store name is a store name in the first target area whose similarity to the store name of the first store meets a first preset condition; based on the store address of the first store, a real store address database and the store name of the first store, searching for a second target store name in a second target area in the electronic map, where the second target store name is a store name in the second target area whose similarity to the store name of the first store meets a second preset condition; when the first target store name or the second target store name exists, determining that the first store is a real store. According to the embodiments of the present application, the efficiency of verifying the authenticity of a store is improved.

Description

Technical Field

[0001] This application belongs to the field of data processing, and particularly relates to a method, apparatus, device, and storage medium for verifying real stores. Background Art

[0002] With the development of electronic payment technology, the application scope of electronic payment has become increasingly wide. When people make payments in stores, they mostly choose electronic payment in most cases. In some cases, some people will forge stores and use the forged stores to collect money, posing risks to users' electronic payments. Therefore, determining the authenticity of stores has become one of the key concerns.

[0003] At present, to determine the authenticity of a store, it is necessary to dispatch specialized staff for grid inspections, that is, dispatch staff to conduct on-site inspections one by one at the location of the store. However, the speed of manually inspecting and determining the authenticity of a store is slow and the efficiency is very low. Summary of the Invention

[0004] The embodiments of this application provide a method, apparatus, device, and storage medium for verifying real stores, which can improve the efficiency of verifying the authenticity of stores.

[0005] In a first aspect, the embodiments of this application provide a method for verifying a real store, including: obtaining the store information of a first store as input, where the store information includes the store name and the store address; based on the store name of the first store and the store address of the first store, searching for a first target store name in a first target area in an electronic map, where the first target store name is a store name in the first target area whose similarity to the store name of the first store meets a first preset condition; based on the store address of the first store, a preset real store address database, and the store name of the first store, searching for a second target store name in a second target area in the electronic map, where the second target store name is a store name in the second target area whose similarity to the store name of the first store meets a second preset condition, and the real store address database includes the store addresses of real stores; in the case where the first target store name or the second target store name exists, determining that the first store is a real store.

[0006] Second aspect, an embodiment of the present application provides a verification device for a real store, including: an acquisition module, configured to acquire the store information of a first store input, where the store information includes the store name and the store address; a first search module, configured to search for a first target store name in a first target area in an electronic map based on the store name of the first store and the store address of the first store, where the first target store name is a store name in the first target area whose similarity to the store name of the first store meets a first preset condition; a second search module, configured to search for a second target store name in a second target area in the electronic map based on the store address of the first store, a preset real store address database, and the store name of the first store, where the second target store name is a store name in the second target area whose similarity to the store name of the first store meets a second preset condition, and the real store address database includes the store addresses of real stores; a determination module, configured to determine that the first store is a real store when the first target store name or the second target store name exists.

[0007] Third aspect, an embodiment of the present application provides a verification device for a real store, including: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the verification method for a real store in the first aspect is implemented.

[0008] Fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the verification method for a real store in the first aspect is implemented.

[0009] An embodiment of the present application provides a verification method, device, device, and storage medium for a real store. In the embodiment of the present application, based on the store name and store address of the input first store, a store name whose similarity to the store name of the first store meets the condition, that is, the first target store name, can be searched in the first target area in the electronic map. It is also possible to concurrently search for a store name whose similarity to the store name of the first store meets the condition, that is, the second target store name, in the second target area in the electronic sub-map based on the store address of the input first store, the store name of the input first store, and the store addresses of real stores in the preset real store address database. Being able to find at least one of the first target store name and the second target name indicates that there is a physical store of the first store in the electronic map, that is, there is a physical store in the real world, so that the authenticity of the store can be determined. This process does not require manual participation and improves the efficiency of verifying the authenticity of the store. Description of the Drawings

[0010] To more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required for use in the embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0011] Figure 1 It is a flowchart of an embodiment of the method for verifying a real store provided by the present application;

[0012] Figure 2 It is a flowchart of another embodiment of the method for verifying a real store provided by the present application;

[0013] Figure 3 It is a schematic diagram of an example of a first region and a second region provided by the embodiments of the present application;

[0014] Figure 4 It is a flowchart of still another embodiment of the method for verifying a real store provided by the present application;

[0015] Figure 5 It is a schematic structural diagram of an embodiment of the device for verifying a real store provided by the present application;

[0016] Figure 6 It is a schematic structural diagram of another embodiment of the device for verifying a real store provided by the present application;

[0017] Figure 7 It is a schematic structural diagram of an embodiment of the device for verifying a real store provided by the present application. Detailed Description of the Embodiments

[0018] The following will describe in detail the features and exemplary embodiments of various aspects of the present application. To make the objectives, technical solutions, and advantages of the present application more clear and understandable, the following further describes the present application in detail in combination with the drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application and not to limit the present application. For those skilled in the art, the present application can be implemented without some of these specific details. The following description of the embodiments is only intended to provide a better understanding of the present application by showing examples of the present application.

[0019] With the development of electronic payment technology, the application scope of electronic payment has become increasingly extensive. When people make payments in stores, they mostly choose electronic payment. In some cases, some people will forge stores and use the forged stores to collect money, which brings risks to users' electronic payments. Therefore, judging the authenticity of stores has become one of the key issues of concern. A real store should correspond to a physical store in the real world. To judge the authenticity of a store, special staff can be dispatched for grid inspections, that is, sending staff to conduct on-site inspections one by one at the location of the store. However, the speed of judging the authenticity of a store by manual inspection is slow and the efficiency is very low. Moreover, due to the uneven qualifications of the staff, there may be misjudgments and wrong judgments, which will also affect the accuracy of judging the authenticity of the store.

[0020] The embodiments of the present application provide a method, device, equipment and storage medium for verifying real stores. It can combine with an electronic map and use the similarity comparison between the input store information and the store information in the electronic map to verify the authenticity of the store. There is no need for manual inspection, which can improve the efficiency of verifying the authenticity of the store and, to a certain extent, also improve the accuracy of verifying the authenticity of the store.

[0021] The method, device, equipment and storage medium for verifying real stores in the embodiments of the present application will be described one by one below.

[0022] The first aspect of the present application provides a method for verifying real stores, which can be applied to a verification device or equipment for real stores, that is, the method for verifying real stores can be executed by a verification device or equipment for real stores, which is not limited here. Figure 1 It is a flowchart of an embodiment of the method for verifying real stores provided by the present application. As Figure 1 shown, the method for verifying real stores may include steps S101 to S104.

[0023] In step S101, obtain the store information of the input first store.

[0024] The first store is the store to be verified. The store information includes the store name and the store address. The store information of the first store can be in text form, or in text form converted from a picture or other forms. For example, the store name of the first store is "A3 Co., Ltd. in A2 City, A1 Province", and the store address of the first store is "No. A6, A5 Street, A4 District, A2 City, A1 Province".

[0025] In some examples, redundancy removal can be performed on the store name. Redundancy removal is to remove the fields that have an adverse effect on the accuracy of similarity from the store name. For example, fields such as "Co., Ltd.", "Company", "XX Province", "XX City", "Self-employed" are frequently occurring fields in the collected store names, that is, redundant fields, which are likely to have an adverse effect on similarity. Therefore, the redundant fields in the store name can be removed in advance, and the subsequent steps can be carried out using the store name after removing the redundant fields.

[0026] In step S102, based on the store name of the first store and the store address of the first store, search for the first target store name in the first target area in the electronic map.

[0027] The first target store name is the store name in the first target area whose similarity to the store name of the first store meets the first preset condition. Specifically, the search function of the electronic map can be used to search for the location obtained by searching the input store address of the first store in the electronic map, and search for the store name in an area including this location, that is, the first target area, whose similarity to the store name of the first store meets the first preset condition, and use the store name that meets the first preset condition as the first target store name.

[0028] The first preset condition is used to screen the store name in the first target area that is most likely to indicate the first store, and there is no limitation here. In some examples, the first preset condition may include that the similarity between the store name in the first target area and the store name of the first store is the highest and greater than the true similarity threshold. That is to say, the first preset condition may include two conditions. Condition 1: The first target store name is the store name in the first target area with the highest similarity to the store name of the first store; Condition 2: The similarity between the first target store name and the store name of the first store is greater than the true similarity threshold. The true similarity threshold can be set according to scenarios, requirements, experience, etc., and there is no limitation here. For example, the true similarity threshold can be 0.6.

[0029] However, it should be noted that there may also be a situation where no store name whose similarity to the store name of the first store meets the first preset condition can be found in the first target area, that is, the first target store name may not exist.

[0030] In step S103, based on the store address of the first store, the preset true store address database, and the store name of the first store, search for the second target store name in the second target area in the electronic map.

[0031] The real store address database includes the store addresses of real stores. The real store address database may include the store addresses of real stores accumulated through various methods. The second target store name is the store name in the second target area whose similarity to the store name of the first store meets the second preset condition. Specifically, the store address of the real store most similar to the store address of the first store can be found in the real store address database first, and then using the search function of the electronic map, search for the location obtained by searching for the store address of the most similar real store in the electronic map. In a region including this location, that is, the second target area, search for the store name whose similarity to the store name of the first store meets the second preset condition, and use the store name that meets the second preset condition as the second target store name.

[0032] The second preset condition is used to screen the store name in the second target area that is most likely to indicate the first store, and is not limited here. In some examples, the second preset condition may include that the similarity between the store name in the second target area and the store name of the first store is the highest and greater than the real similarity threshold. That is to say, the second preset condition may include two conditions. Condition 1: The second target store name is the store name in the second target area with the highest similarity to the store name of the first store. Condition 2: The similarity between the second target store name and the store name of the first store is greater than the real similarity threshold. For the specific content of the real similarity threshold, reference can be made to the relevant description in the above embodiments and will not be elaborated here.

[0033] However, it should be noted that there may also be a situation where no store name whose similarity to the store name of the first store meets the second preset condition can be found in the second target area, that is, the second target store name may not exist.

[0034] In step S104, when the first target store name or the second target store name exists, it is determined that the first store is a real store.

[0035] The existence of at least one of the first target store name and the second target store name indicates that the first store exists in the area in the electronic map that matches the description of the store information of the first store, that is, the first store has a physical store in the real world, and it can be determined that the first store is a real store.

[0036] In an embodiment of the present application, based on the store name of the first store and the store address of the first store input, a store name that meets the condition of similarity to the store name of the first store, that is, the first target store name, can be searched in the first target area in the electronic map. Additionally, in parallel, based on the store address of the first store input, the store name of the first store input, and the store addresses of real stores in the preset real store address database, a store name that meets the condition of similarity to the store name of the first store, that is, the second target store name, can be searched in the second target area in the electronic sub-map. Being able to find at least one of the first target store name and the second target name indicates that there is a physical store of the first store in the electronic map, that is, there is a physical store in the real world, thereby the authenticity of the store can be determined. This process does not require manual participation, improving the efficiency of verifying the authenticity of the store. Moreover, by using the input store information in combination with the electronic map to search for similar store names in two target areas based on different conditions in the electronic map, errors in the search process can also be avoided, improving the accuracy of verifying the authenticity of the store.

[0037] In some embodiments, the electronic map can be used to respectively convert the store name of the first store and the store address of the first store into corresponding geographical locations, thereby obtaining the first target area and the second target area. Figure 2 It is a flowchart of another embodiment of the method for verifying real stores provided by the present application. Figure 2 Different from Figure 1 is that Figure 1 step S102 in Figure 2 can be specifically refined into Figure 1 steps S1021 to S1025 in Figure 2 and step S103 in

[0038] In step S1021, using the electronic map, the store address of the first store is converted into a first geographical location.

[0039] Specifically, the interface for searching for surrounding Points of Interest (POIs) in the electronic map can be called, and the store address of the first store is input into this interface of the electronic map, so as to use the electronic map to convert the store address of the first store into a geographical location, that is, the first geographical location. The first geographical location is the geographical location obtained by converting the store address of the first store using the electronic map. The first geographical location may specifically include longitude and latitude, or may include other specific parameters, which are not limited herein.

[0040] In step S1022, the similarity between the store name of the first store and the store names of each store in the first area where the first geographical location is located in the electronic map is calculated.

[0041] The first target area includes a first area. The first area includes a first geographical location, and the size, shape, and position of the first area can be set according to scenarios, requirements, experience, etc., and are not limited herein. In some examples, the first area can be a circular area centered on the first geographical location with a radius of r1, and the distance r1 can be set according to scenarios, requirements, experience, etc. For example, the distance r1 can be 1000 meters.

[0042] The higher the similarity between the store name of the first store and the store name of a certain store in the first area, the higher the possibility that the first store and this store in the first area are the same store. The similarity between the store name of the first store and the store names of each store in the first area can be calculated based on the similarity in aspects of the store name attributes. The aspects of the store name attributes can include one or more of semantic aspects, feature structure aspects, word order aspects, etc., and are not limited herein.

[0043] In step S1023, according to the similarity between the store name of the first store and the store names of each store in the first area and the first preset condition, search for the first target store name.

[0044] If there is a store in the first area whose store name similarity with the store name of the first store meets the first preset condition, the store name of this store in the first area can be used as the first target store name.

[0045] For example, the first preset condition includes: the similarity with the store name of the first store in the first target area is the highest and greater than the true similarity threshold. There are 5 stores in the first area, and the store names of the 5 stores are XXXX1, XXXX2, XXXX3, XXXX4, and XXXX5 respectively. Among them, the similarity between XXXX3 and the store name of the first store is the highest among the similarities between the store names of these 5 stores and the store name of the first store, and the similarity between XXXX3 and the store name of the first store is greater than the true similarity threshold, then XXXX3 can be used as the first target store name.

[0046] In step S1024, if the first target store name is not found in the first area, calculate the similarity between the store name of the first store and the store names of each store in the second area surrounding the first area in the electronic map.

[0047] In some cases, due to errors in the conversion of the store name of the first store to the first geographical location by the electronic map or other factors, the store name of the store whose similarity to the store name of the first store meets the first preset condition cannot be found within the first area, that is, the first target store name cannot be found within the first area. In this case, the scope of the first target area can be expanded, and the first target store name can be searched for continuously within the expanded scope.

[0048] Specifically, the first target area may further include a second area. The second area is an extension outward from the first area, and the second area can surround the first area. In some examples, the inner boundary of the second area may coincide with or be adjacent to the outer boundary of the first area, and the distance from the outer boundary of the first area to the first geographical location is less than the distance from the outer boundary of the second area to the first geographical location. The second area can be set around the first geographical location, and the distance from the outer boundary of the second area to the first geographical location can be determined according to scenarios, requirements, experience, etc., and is not limited here. For example, the distance from the outer boundary of the second area to the first geographical location can be 2000 meters. For example, Figure 3 is a schematic diagram of an example of the first area and the second area provided by the embodiment of the present application. As Figure 3 shown, the first geographical location is O(gps_x, gps_y), where gps_x is the longitude and gps_y is the latitude; the first area is a circular area centered on the first geographical location with a radius of r1; the inner boundary of the second area coincides with the outer boundary of the first area, and the distance from the outer boundary of the second area to the first geographical location O is r2, and the second area is Figure 3 the annular shaded area in.

[0049] The higher the similarity between the store name of the first store and the store name of a certain store in the second area, the higher the possibility that the first store and this store in the second area are the same store. The similarity between the store name of the first store and the store names of each store in the second area can be calculated based on the similarity in the attributes of the store names. The attribute aspects of the store name can include one or more of semantic aspects, feature structure aspects, word order aspects, etc., and are not limited here.

[0050] In step S1025, according to the similarity between the store name of the first store and the store names of each store in the second area and the first preset condition, the first target store name is searched for.

[0051] In the case where the similarity between the store name of a store in the second area and the store name of the first store meets the first preset condition, the store name of this store in the second area can be used as the first target store name.

[0052] In step S1031, from a preset real store address database, find the first store address with the highest similarity to the store address of the first store.

[0053] Specifically, the similarity between the store address of the first store and the store addresses of each real store in the real store address database can be calculated, and the store address of the real store with the highest similarity is determined as the first store address. The similarity between the store address of the first store and the store addresses of the real stores can be calculated based on the similarity in terms of the attributes of the store address. The aspects of the attributes of the store address can include one or more of semantic aspects, feature structure aspects, word order aspects, etc., and are not limited herein.

[0054] In step S1032, use an electronic map to convert the first store address into a second geographical location.

[0055] Specifically, the interface of the electronic map for searching for surrounding points of interest can be called, the first store address is input into this interface of the electronic map, and using the electronic map, the first store address is converted into a geographical location, that is, the second geographical location. The second geographical location is the geographical location obtained by converting the first store address using the electronic map. The second geographical location can specifically include longitude and latitude, and can also include other specific parameters, and is not limited herein.

[0056] In step S1033, calculate the similarity between the store name of the first store and the store names of each store in the third area where the second geographical location is located in the electronic map.

[0057] The second target area includes the third area. The third area includes the second geographical location. The size, shape, and location of the third area can be set according to scenarios, requirements, experience, etc., and are not limited herein. In some examples, the third area can be a circular area centered on the second geographical location with a radius of r3, and the distance r3 can be set according to scenarios, requirements, experience, etc. For example, the distance r3 can be 1000 meters.

[0058] The higher the similarity between the store name of the first store and the store name of a certain store in the third area, the higher the possibility that the first store and this store in the third area are the same store. The similarity between the store name of the first store and the store names of each store in the third area can be calculated based on the similarity in terms of the attributes of the store name. The aspects of the attributes of the store name can include one or more of semantic aspects, feature structure aspects, word order aspects, etc., and are not limited herein.

[0059] In step S1034, according to the similarity between the store name of the first store and the store names of each store in the third area and the second preset condition, find the second target store name.

[0060] When the similarity between the store name of a store in the third region and the store name of the first store meets the second preset condition, the store name of this store in the third region can be used as the second target store name.

[0061] In step S1035, when the second target store name cannot be found in the third region, calculate the similarity between the store name of the first store and the store names of each store in the fourth region that surrounds the third region.

[0062] In some cases, due to errors in converting geographical locations on the electronic map, similarity calculation errors, or other factors, the store name of a store whose similarity to the store name of the first store meets the second preset condition cannot be found in the third region, that is, the second target store name cannot be found in the third region. In this case, the range of the second target region can be expanded, and the second target store name can be searched for continuously within the expanded range.

[0063] Specifically, the second target region may further include the fourth region. The fourth region is an extension outward from the third region, and the fourth region can surround the third region. In some examples, the inner boundary of the fourth region may coincide with or be adjacent to the outer boundary of the third region, and the distance from the outer boundary of the third region to the second geographical location is less than the distance from the outer boundary of the fourth region to the second geographical location. The fourth region can be set around the second geographical location, and the distance from the outer boundary of the fourth region to the second geographical location can be determined according to scenarios, requirements, experience, etc., and is not limited here. For example, the distance from the outer boundary of the fourth region to the second geographical location can be 2000 meters. The relationship between the third region and the fourth region can refer to the relationship between the first region and the second region in the above embodiments, and will not be elaborated here.

[0064] The higher the similarity between the store name of the first store and the store name of a certain store in the fourth region, the higher the possibility that the first store and this store in the fourth region are the same store. The similarity between the store name of the first store and the store names of each store in the fourth region can be calculated based on the similarity in terms of the attributes of the store name. The aspects of the attributes of the store name can include one or more of semantic aspects, feature structure aspects, word order aspects, etc., and are not limited here.

[0065] In step S1036, according to the similarity between the store name of the first store and the store names of each store in the fourth region and the second preset condition, search for the second target store name.

[0066] When the similarity between the store name of a store in the fourth region and the store name of the first store meets the second preset condition, the store name of this store in the fourth region can be used as the second target store name.

[0067] The above steps S1021 to S1025 are to obtain the first geographical location based on the store name of the first store, and thus select the first target area for searching for the first target store name according to the first geographical location. The above steps S1031 to S1036 are to obtain the second geographical location based on the first store address that is most similar to the store address of the first store in the real store address database, and thus select the second target area for searching for the second target store name according to the second geographical location. Through the mutual complement of the two target areas, namely the first target area and the second target area, the misjudgment of real stores caused by errors in certain aspects is avoided, and the accuracy of store authenticity determination is further improved.

[0068] In some embodiments, it may occur that the first target store name and the second target store name cannot be found. To avoid misjudgment of real stores, the search range can be expanded to the entire area of the electronic map. Figure 4 It is a flowchart of another embodiment of the method for verifying real stores provided by this application. Figure 4 Different from Figure 1 that, Figure 4 the method for verifying real stores shown in

[0069] may further include steps S105 to S109.

[0070] In step S105, in the case where both the first target store name and the second target store name do not exist, calculate the similarity between the store name of the first store and the store names of each store in the electronic map.

[0071] In the case where the first target store name and the second target store name cannot be found, the search range can be continuously expanded. The similarity between the store name of the first store and the store names of each store in the electronic map can be calculated based on the similarity in terms of store name attributes. The aspects of store name attributes can include one or more of semantic aspects, feature structure aspects, word order aspects, etc., which are not limited here.

[0072] In step S106, obtain the candidate stores in the electronic map.

[0073] The candidate stores include stores whose similarity between the store name and the store name of the first store is higher than the real similarity threshold. The specific content of the real similarity threshold can refer to the relevant description in the above embodiments, which is not limited here.

[0074] In step S107, use the electronic map to obtain the geographical locations of the candidate stores.

[0075] In step S108, when the distance between the geographical location of the candidate store and the first geographical location is less than or equal to the preset distance threshold, the first store is determined to be a real store.

[0076] The first geographical location is the geographical location obtained by converting the store address of the first store using an electronic map. The preset distance threshold is used to determine whether the first store is a candidate store. The preset distance threshold can be set according to scenarios, requirements, experience, etc., and is not limited here. For example, the preset distance threshold can be 1000 meters. That the distance between the geographical location of the candidate store and the first geographical location is less than or equal to the preset distance threshold means that the first store is a candidate store. The candidate store is an entity store recorded in the electronic map, so the first store is determined to be a real store.

[0077] In step S109, when the distance between the geographical location of the candidate store and the first geographical location is greater than the preset distance threshold, the first store is determined to be a non-real store.

[0078] That the distance between the geographical location of the candidate store and the first geographical location is greater than the preset distance threshold means that the first store is not a candidate store. No entity store similar enough to the first store is found in the electronic map, so the first store is determined to be a non-real store, that is, it is determined that the first store is not a real store.

[0079] The following specifically describes the calculation of the similarity of the store name and the calculation of the similarity of the store address in the above embodiments.

[0080] The similarity in the above embodiments can be calculated based on the attribute similarity. Specifically, the similarity of the store name can be calculated based on the attribute similarity of the store name. Similarly, the similarity of the store address can be calculated based on the attribute similarity of the store address.

[0081] The attribute similarity is the similarity in terms of the attributes of the objects participating in the similarity calculation. In some examples, the attribute similarity may include, but is not limited to, one or more of the following: semantic similarity, feature structure similarity, word order similarity. For example, the attribute similarity includes semantic similarity, feature structure similarity, and word order similarity.

[0082] The semantic similarity between the first object and the second object refers to the similarity between the first object and the second object in terms of semantics, that is, the meaning of the language. The feature structure similarity between the first object and the second object refers to the similarity between the first object and the second object in terms of language structure. The word order similarity between the first object and the second object refers to the similarity between the first object and the second object in terms of word order, that is, the combination order of words in the language.

[0083] The semantic similarity can be calculated based on the feature vectors obtained by converting the first object and the second object. When the first object includes the store name of the first store, the second object includes store names, and specifically, the second object can include the store names of the stores in the first target area, the store names of the stores in the second target area, the store names of the stores in the electronic map, etc. in the above embodiments, which can be selected according to specific requirements and are not limited herein. When the first object includes the store name of the first store, the second object includes store names, and specifically, the second object can include the store addresses of the real stores in the real store address database, which are not limited herein.

[0084] Specifically, the first object and the second object can be respectively converted into feature vectors through a text feature extraction model. The type of the text feature extraction model is not limited herein. For example, the text feature extraction model can be a BERT (Bidirectional Encoder Representation from Transformers) model. The feature vectors obtained by converting the first object and the second object using the BERT model can be shown as the following relational expressions (1) and (2):

[0085] U = BERT(Mchnt1)[′pooler_output′] (1)

[0086] V = BERT(Mchnt2)[′pooler_output′] (2)

[0087] Where BERT is the BERT model, Mchnt1 is the first object, Mchnt2 is the second object, U is the feature vector obtained by converting the first object, V is the feature vector obtained by converting the second object, and pooler_output is the sentence-level feature vector extracted. The feature vectors obtained by converting the first object and the second object are sentence-level feature vectors.

[0088] For the convenience of processing, the feature vectors obtained by converting the first object and the second object can both be set as d-dimensional feature vectors, where d is a positive integer and can be obtained according to scenarios, requirements, and experience, which are not limited herein. For example, d can be set to 3.

[0089] In some examples, the semantic similarity can be calculated based on the average semantic information of the feature vector of the first object, the average semantic information of the feature vector of the second object, and a preset smoothing coefficient. Based on the feature vector obtained by converting the first object and the feature vector obtained by converting the second object, the first norm of the feature vector obtained by converting the first object and the second norm of the feature vector obtained by converting the second object can be calculated. The first norm can be the average semantic information of the feature vector obtained by converting the first object, and the second norm can be the average semantic information of the feature vector obtained by converting the second object. Then, according to the product of the first norm and the second norm, the square value of the first norm, the square value of the second norm, and a preset first smoothing coefficient, the semantic similarity between the first object and the second object can be calculated.

[0090] For example, the semantic similarity between the first object and the second object can be represented by the feature vector obtained by converting the first object and the feature vector obtained by converting the second object, and can be calculated according to the following relational expressions (3) to (5):

[0091]

[0092]

[0093]

[0094] where U i is the i-th dimensional data in the feature vector obtained by converting the first object; V i is the i-th dimensional data in the feature vector obtained by converting the second object; d is the dimension of the feature vector; μ U is the average semantic information of the feature vector obtained by converting the first object, that is, the norm of the sentence-level feature vector of the first object, which is also the first norm; μ V is the average semantic information of the feature vector obtained by converting the second object, that is, the norm of the sentence-level feature vector of the second object, which is also the second norm; C1 is the first smoothing coefficient; l(U, V) is the semantic similarity between the feature vector obtained by converting the first object and the feature vector obtained by converting the second object. The value of the first smoothing coefficient C1 can be determined according to the scenario, requirements, and experience, and is not limited here. In some examples, the first smoothing coefficient C1 < 1.

[0095] The feature structure similarity can be calculated based on the feature vectors obtained by converting the first object and the second object. When the first object includes the store name of the first store, the second object includes store names, and specifically, the second object can include the store names of the stores in the first target area, the store names of the stores in the second target area, the store names of the stores in the electronic map, etc. in the above embodiments, which can be selected according to specific requirements and are not limited herein. When the first object includes the store name of the first store, the second object includes store names, and specifically, the second object can include the store addresses of the actual stores in the actual store address database, which are not limited herein.

[0096] Specifically, the first object and the second object can be respectively converted into feature vectors through a text feature extraction model. The content of the text feature extraction model and the feature vectors can be referred to the relevant descriptions in the above embodiments and will not be elaborated herein. The relational expressions of the feature vector obtained by converting the first object and the feature vector obtained by converting the second object can be referred to the relational expressions (1) and (2) in the above embodiments.

[0097] In some examples, the feature structure similarity can be calculated based on the standard deviation of the feature vector of the first object, the standard deviation of the feature vector of the second object, the covariance of the feature vector of the first object and the feature vector of the second object, and a preset second smoothing coefficient. Based on the feature vector obtained by converting the first object and the feature vector obtained by converting the second object, the first standard deviation of the feature vector obtained by converting the first object, the second standard deviation of the feature vector obtained by converting the second object, and the covariance of the feature vector obtained by converting the first object and the feature vector obtained by converting the second object are calculated. Then, according to the product of the first standard deviation and the second standard deviation, the covariance, and the preset second smoothing coefficient, the feature structure similarity between the first object and the second object is calculated. It should be noted that the second smoothing coefficient used to calculate the feature structure similarity and the first smoothing coefficient used to calculate the semantic similarity can be the same or different, which are not limited herein.

[0098] For example, the semantic similarity between the first object and the second object can be represented by the feature structure similarity between the feature vector obtained by converting the first object and the feature vector obtained by converting the second object, and can be calculated according to the following relational expressions (6) to (9):

[0099]

[0100]

[0101]

[0102]

[0103] Among them, U i 、Vi , μ U , μ V The definitions of μ and d are the same as those in the above embodiments and will not be elaborated here; σ U is the standard deviation of the feature vector obtained by converting the first object; σ V is the standard deviation of the feature vector obtained by converting the second object; σ UV is the covariance of the feature vector obtained by converting the first object and the feature vector obtained by converting the second object, that is, the degree of consistency of the sentence-level features changing in each semantic direction; s(U, V) is the feature structure similarity between the feature vector obtained by converting the first object and the feature vector obtained by converting the second object; C2 is a preset second smoothing coefficient, which can be set according to scenarios, requirements, experience, etc., and is not limited here. For example, the second smoothing coefficient C2 < 1.

[0104] The word order similarity is calculated based on the feature matrix obtained by converting the first object and the feature matrix obtained by converting the second object. When the first object includes the store name of the first store, the second object includes the store name, and the second object may specifically include the store names of the stores in the first target area, the store names of the stores in the second target area, the store names of the stores in the electronic map, etc. in the above embodiments, which can be selected according to specific requirements and are not limited here. When the first object includes the store name of the first store, the second object includes the store name, and the second object may specifically include the store addresses of the real stores in the real store address database, which is not limited here.

[0105] Specifically, through the text feature extraction model, the first object and the second object can be respectively converted into feature matrices. The type of the text feature extraction model is not limited here. For example, the text feature extraction model can be a BERT model. The feature matrices obtained by converting the first object and the second object using the BERT model can be shown in the following relational expressions (10) and (11):

[0106] N = BERT(Mchnt1)[′last_hidden_state′] (10)

[0107] M = BERT(Mchnt2)[′last_hidden_state′] (11)

[0108] Among them, the definitions of BERT, Mchnt1, and Mchnt2 are the same as those in the above embodiments and will not be elaborated here; N is the feature matrix obtained by converting the first object; M is the feature matrix obtained by converting the second object; last_hidden_state is the word-level feature matrix extracted. The feature matrices obtained by converting the first object and the second object are word-level feature matrices.

[0109] For ease of processing, the feature matrix obtained by converting the first object can be set as a feature matrix with a length equal to the text length n of the first object and a width equal to the dimension d, that is, the feature matrix obtained by converting the first object is an n×d matrix. The feature matrix obtained by converting the second object can be set as a feature matrix with a length equal to the text length m of the second object and a width equal to the dimension d, and the feature matrix obtained by converting the second object is an m×d matrix. The dimension d of the feature matrix can be the same as the dimension d of the feature vector in the above embodiments.

[0110] In some examples, the word order similarity can be calculated using an N-gram model. The word order similarity can be calculated based on the similarity between consecutive multiple word vectors in the feature matrix obtained by converting the first object and consecutive multiple word vectors in the feature matrix obtained by converting the second object, the text length of the first object, the text length of the second object, and a penalty term for the text lengths of the first object and the second object. The first word vector formed by obtaining column elements can be acquired from the feature matrix obtained by converting the first object. The second word vector formed by obtaining column elements can be acquired from the feature matrix obtained by converting the second object. Each column element in the feature matrix obtained by converting the first object can form a first word vector. Each column element in the feature matrix obtained by converting the second object can form a second word vector. According to the first word vector and the second word vector, calculate the word vector similarity between consecutive x first word vectors and consecutive x second word vectors, where x = 1, 2, 3, …. The similarity between consecutive x first word vectors and consecutive x second word vectors can represent the similarity between a substring with a text length of x in the first object and a substring with a text length of x in the second object. According to multiple word vector similarities, the text length of the first object, the text length of the second object, and a penalty term for the text lengths of the first object and the second object, the word order similarity is calculated. Combine the similarities of substrings with a text length of x in the first object and substrings with a text length of x in the second object when x takes different values to obtain the word order similarity between the first object and the second object.

[0111] For example, the word order similarity between the first object and the second object can be represented by the word order similarity between the feature matrix obtained by converting the first object and the feature matrix obtained by converting the second object, and can be calculated according to the following relational expressions (12) to (14):

[0112]

[0113]

[0114]

[0115] where max is for calculating the maximum value; min is for calculating the minimum value; P iis the matching degree under the condition that the length of the matching text is i, that is, the matching degree of the i-gram; n is the text length of the first object; m is the text length of the second object; N j is the first word vector formed by the elements in the j-th column of the feature matrix obtained by converting the first object; M j+k is the second word vector formed by the elements in the (j + k)-th column of the feature matrix obtained by converting the second object; is the dot product calculation; is the dot product calculation of the word-level features of the feature matrix obtained by converting the first object and the feature matrix obtained by converting the second object, indicating the similarity between the two; is the sum of the similarities calculated for m - i rounds; P i is to take the average of the sum of the similarities calculated for m - i rounds. is the similarity between k consecutive first word vectors in the feature matrix obtained by converting the first object and k consecutive second word vectors in the feature matrix obtained by converting the second object; is the similarity between k - 1 consecutive first word vectors in the feature matrix obtained by converting the first object and k - 1 consecutive second word vectors in the feature matrix obtained by converting the second object; is the similarity between k - n + i consecutive first word vectors in the feature matrix obtained by converting the first object and k - n + i consecutive second word vectors in the feature matrix obtained by converting the second object; penalty is the penalty term for the text lengths of the first object and the second object; o(N, M) is the word order similarity between the feature matrix obtained by converting the first object and the feature matrix obtained by converting the second object; ω i is the reciprocal of i. The maximum value of i is not limited here and can be set according to scenarios, requirements, experience, etc. In some examples, for the convenience of calculation, the maximum value of i can be made less than or equal to 4. In the relational expressions (12) to (14), the greater the difference between the text lengths of the first object and the second object, the lower the value of the penalty term, and the lower the word order similarity between the first object and the second object.

[0116] In some examples, the similarity between the first object and the second object can be the a i power of the product of the similarities of each attribute. a i is the weight parameter of the attribute similarity described in the i-th item, and i is a positive integer. The weight parameter can be set according to the importance of each attribute similarity and is not limited here. In the case where the similarity includes one attribute similarity, the similarity between the first object and the second object can be the a1 power of this one attribute similarity. It should be noted that a i the i in is independent of the i in the above relational expressions (1) to (14) and does not affect each other.

[0117] For example, if the attribute similarity includes semantic similarity, the similarity calculation can be obtained according to the following relational expression (15):

[0118]

[0119] Among them, l(U, V) and a1 can be referred to the definitions in the above embodiments and will not be elaborated here; sim(Mchnt1, Mchnt2) is the similarity between the first object and the second object.

[0120] Again, for example, if the attribute similarity includes semantic similarity and feature structure similarity, the similarity calculation can be obtained according to the following relational expression (16):

[0121]

[0122] Among them, sim(Mchnt1, Mchnt2), l(U, V), s(U, V), a1 and a2 can be referred to the definitions in the above embodiments and will not be elaborated here.

[0123] Once again, for example, if the attribute similarity includes semantic similarity, feature structure similarity and word order similarity, the similarity calculation can be obtained according to the following relational expression (17):

[0124]

[0125] Among them, sim(Mchnt1, Mchnt2), l(U, V), s(U, V), o(U, V), a1, a2 and a3 can be referred to the definitions in the above embodiments and will not be elaborated here.

[0126] Next, an example is used to illustrate the above verification method for real stores, where the real similarity threshold is 0.6.

[0127] Input the store name "XXX" of the first store, and this store name "XXX" does not include redundant fields. Input the store name "XXX" into the electronic map, and through the conversion of the electronic map, the first geographical location (gps_x1, gps_y1) is obtained. The store names that can be found within 1000 meters around the first geographical location (gps_x1, gps_y1) include "YYY1", "YYY2" and "YYY3". Among them, the similarity between the store name "XXX" and the store name "YYY1" is the highest. The following gives the similarity calculation process between the store name "XXX" and the store name "YYY1".

[0128] The dimension of the feature vector and the dimension d of the feature matrix are 3. The attribute similarity includes semantic similarity, feature structure similarity and word order similarity, and a1 = a2 = a3 = 1. According to the above relational expressions (1) to (14) and (17), the following calculation formula can be obtained:

[0129] U = [-0.1170 -0.0428 -0.1075];

[0130] V = [0.0212 -0.2200 -0.2130];

[0131]

[0132]

[0133]

[0134]

[0135] l(U, V) = 0.9164;

[0136] σ U = 0.0404;

[0137] σ V = 0.1373;

[0138] σ UV = 0.003;

[0139] s(U, V) = 0.6110;

[0140]

[0141] P2 ≈ P3 ≈ 0;

[0142] penalty ≈ 1;

[0143] o(N, M) ≈ 0.8181;

[0144] s(U, V) = 0.6110;

[0145] sim(Mchnt1, Mchnt2) = 0.9146 × 0.6110 × 0.8181 = 0.4581;

[0146] The similarity between the store name "XXX" and the store name "YYY1" is 0.4581. This similarity is less than the true similarity threshold of 0.6. Therefore, the first target store name was not found within 1000 meters around the first geographical location (gps_x1, gps_y1). The search range was expanded, and the search and similarity calculation were performed again within 1000 to 2000 meters around the first geographical location (gps_x1, gps_y1). The first target store name was not found within 1000 to 2000 meters around the first geographical location (gps_x1, gps_y1).

[0147] From the preset real store address database, the first store address "ZZZZ2" with the highest similarity to the store address "ZZZZ1" of the first store is found, and the first store address "ZZZZ2" is input into the electronic map to obtain the second geographical location (gps_x2, gps_y3) after conversion. Search for the store name "XXX" with the highest similarity to the store name "XXX" within 1000 meters around the second geographical location (gps_x2, gps_y3). The similarity between the found store name "XXX" and the store name "XXX" of the first store is greater than the real similarity threshold, that is, there is a second target store name, and it can be determined that the first store with the input store name "XXX" is a real store.

[0148] The second aspect of the present application provides a verification device for real stores. Figure 5 It is a schematic structural diagram of an embodiment of the verification device for real stores provided by the present application. As Figure 5 shown, the verification device 200 for real stores may include an acquisition module 201, a first search module 202, a second search module 203, and a determination module 204.

[0149] The acquisition module 201 may be configured to acquire the store information of the input first store.

[0150] The store information includes the store name and the store address.

[0151] The first search module 202 may be configured to search for the first target store name in the first target area in the electronic map based on the store name of the first store and the store address of the first store.

[0152] The first target store name is the store name in the first target area whose similarity to the store name of the first store meets the first preset condition.

[0153] In some examples, the first preset condition includes: the highest similarity to the store name of the first store in the first target area and greater than the real similarity threshold.

[0154] The second search module 203 may be configured to search for the second target store name in the second target area in the electronic map based on the store address of the first store, the preset real store address database, and the store name of the first store.

[0155] The real store address database includes the store addresses of real stores. The second target store name is the store name in the second target area whose similarity to the store name of the first store meets the second preset condition.

[0156] In some examples, the second preset condition includes: the highest similarity to the store name of the first store in the second target area and greater than the real similarity threshold.

[0157] The determination module 204 can be used to determine that the first store is a real store when the first target store name or the second target store name exists.

[0158] In the embodiments of the present application, based on the input store name of the first store and the store address of the first store, a store name that meets the condition of similarity to the store name of the first store, that is, the first target store name, can be searched in the first target area in the electronic map. Additionally, in parallel, based on the input store address of the first store, the input store name of the first store, and the store addresses of real stores in the preset real store address database, a store name that meets the condition of similarity to the store name of the first store, that is, the second target store name, can be searched in the second target area in the electronic map. Being able to find at least one of the first target store name and the second target name indicates that the first store has a physical store in the electronic map, that is, a physical store exists in the real world, thereby the authenticity of the store can be determined. This process does not require manual participation, improving the efficiency of verifying the authenticity of the store. Moreover, by using the input store information in combination with the electronic map and searching for similar store names in two target areas based on different conditions in the electronic map, errors in the search process can also be avoided, improving the accuracy of verifying the authenticity of the store.

[0159] In some embodiments, the first search module 202 can be used to: convert the store address of the first store into a first geographical location using the electronic map; calculate the similarity between the store name of the first store and the store names of each store in the first area where the first geographical location is located in the electronic map, and the first target area includes the first area; search for the first target store name according to the similarity between the store name of the first store and the store names of each store in the first area and the first preset condition.

[0160] In some embodiments, the first search module 202 can also be used to: when the first target store name is not found in the first area, calculate the similarity between the store name of the first store and the store names of each store in the second area surrounding the first area in the electronic map, and the first target area further includes the second area; search for the first target store name according to the similarity between the store name of the first store and the store names of each store in the second area and the first preset condition.

[0161] In some examples, the first area is centered on the first geographical location. The inner boundary of the second area coincides with or is adjacent to the outer boundary of the first area, and the distance from the outer boundary of the first area to the first geographical location is less than the distance from the outer boundary of the second area to the first geographical location.

[0162] In some embodiments, the second search module 202 can be used to: search for the first store address with the highest similarity to the store address of the first store from a preset real store address database; use an electronic map to convert the first store address into a second geographical location; calculate the similarity between the store name of the first store and the store names of each store in the third area where the second geographical location is located in the electronic map, and the second target area includes the third area; search for the second target store name according to the similarity between the store name of the first store and the store names of each store in the third area and a second preset condition.

[0163] In some embodiments, the second search module 202 can also be used to: in the case where the second target store name is not found in the third area, calculate the similarity between the store name of the first store and the store names of each store in the fourth area surrounding the third area, and the second target area further includes the fourth area; search for the second target store name according to the similarity between the store name of the first store and the store names of each store in the fourth area and a second preset condition.

[0164] In some examples, the third area is centered on the second geographical location. The inner boundary of the fourth area coincides with or is adjacent to the outer boundary of the third area, and the distance from the outer boundary of the third area to the second geographical location is less than the distance from the outer boundary of the fourth area to the second geographical location.

[0165] Figure 6 It is a schematic structural diagram of another embodiment of the verification device for real stores provided by this application. Figure 6 Different from Figure 5 is that Figure 6 the verification device 200 for real stores shown in

[0166] The third search module 205 can be used to: in the case where both the first target store name and the second target store name do not exist, calculate the similarity between the store name of the first store and the store names of each store in the electronic map; obtain candidate stores in the electronic map, and the candidate stores include stores with a similarity between the store name and the store name of the first store higher than the real similarity threshold; use the electronic map to obtain the geographical locations of the candidate stores.

[0167] The above determination module 204 can also be used to: in the case where the distance between the geographical location of the candidate store and the first geographical location is less than or equal to a preset distance threshold, determine that the first store is a real store, and the first geographical location is the geographical location obtained by converting the store name of the first store using the electronic map; in the case where the distance between the geographical location of the candidate store and the first geographical location is greater than the preset distance threshold, determine that the first store is not a real store.

[0168] The similarity in the above embodiments is calculated based on attribute similarity.

[0169] In some examples, the attribute similarity includes one or more of the following: semantic similarity, feature structure similarity, and word order similarity.

[0170] In some examples, the similarity is the product of the ath powers of the respective attribute similarities, where a i is the weight parameter of the ith attribute similarity. i is the weight parameter of the ith attribute similarity.

[0171] In some examples, the semantic similarity is calculated based on the feature vectors obtained by transforming the first object and the feature vectors obtained by transforming the second object.

[0172] The feature structure similarity is calculated based on the feature vectors obtained by transforming the first object and the feature vectors obtained by transforming the second object.

[0173] The word order similarity is calculated based on the feature matrices obtained by transforming the first object and the feature matrices obtained by transforming the second object.

[0174] Among them, when the first object includes the store name of the first store, the second object includes the store name. When the first object includes the store address of the first store, the second object includes the store address.

[0175] In some examples, the above verification device 200 for real stores may further include a calculation module.

[0176] The calculation module can be used to: based on the feature vectors obtained by transforming the first object and the feature vectors obtained by transforming the second object, calculate the first norm of the feature vectors obtained by transforming the first object and the second norm of the feature vectors obtained by transforming the second object; according to the product of the first norm and the second norm, the square value of the first norm, the square value of the second norm, and a preset first smoothing coefficient, calculate the semantic similarity between the first object and the second object.

[0177] The calculation module can be used to: based on the feature vectors obtained by transforming the first object and the feature vectors obtained by transforming the second object, calculate the first standard deviation of the feature vectors obtained by transforming the first object, the second standard deviation of the feature vectors obtained by transforming the second object, and the covariance between the feature vectors obtained by transforming the first object and the feature vectors obtained by transforming the second object; according to the product of the first standard deviation and the second standard deviation, the covariance, and a preset second smoothing coefficient, calculate the feature structure similarity between the first object and the second object.

[0178] The computing module can be used to: obtain a first word vector formed by column elements from a feature matrix obtained by transforming a first object; obtain a second word vector formed by column elements from a feature matrix obtained by transforming a second object; calculate the word vector similarity between consecutive x first word vectors and consecutive x second word vectors according to the first word vector and the second word vector, where x = 1, 2, 3, …; calculate an order similarity according to multiple word vector similarities, the text length of the first object, the text length of the second object, and a penalty term for the text lengths of the first object and the second object.

[0179] The third aspect of this application also provides a verification device for a real store. Figure 7 It is a schematic structural diagram of an embodiment of the verification device for a real store provided by this application. As Figure 7 shown, the verification device 300 for a real store includes a memory 301, a processor 302, and a computer program stored on the memory 301 and executable on the processor 302.

[0180] In one example, the above-mentioned processor 302 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0181] The memory 301 may include a read-only memory (ROM), a random access memory (RAM), a disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical, or other physical / tangible memory storage device. Thus, generally, the 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 verification method for a real store according to the embodiments of this application.

[0182] The processor 302 runs a computer program corresponding to the executable program code by reading the executable program code stored in the memory 301, so as to implement the verification method for a real store in the above embodiments.

[0183] In one example, the verification device 300 for a real store may further include a communication interface 303 and a bus 304. Among them, as Figure 7 shown, the memory 301, the processor 302, and the communication interface 303 are connected through the bus 304 and complete communication with each other.

[0184] The communication interface 303 is mainly used to implement communication between various modules, devices, units, and / or devices in the embodiments of the present application. The input device and / or output device can also be accessed through the communication interface 303.

[0185] The bus 304 includes hardware, software, or both, and couples the components of the verification device 300 of the real store to each other. By way of example and not limitation, the 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 InfiniBand 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. In a suitable case, the bus 304 may include one or more buses. Although the embodiments of the present application describe and illustrate specific buses, the present application contemplates any suitable bus or interconnect.

[0186] The fourth aspect of the present application further provides a computer-readable storage medium, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the verification method of the real store in the above embodiments can be implemented, and the same technical effects can be achieved. To avoid repetition, it will not be elaborated here. Among them, the above computer-readable storage medium may include a non-transitory computer-readable storage medium, such as a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disc, etc., which is not limited herein.

[0187] It should be clear that each embodiment in this specification is described in a progressive manner. For the parts that are the same or similar among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. For the device embodiments, equipment embodiments, and computer-readable storage medium embodiments, the relevant parts can refer to the description part 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. And, for the sake of brevity, the detailed description of known method technologies is omitted here.

[0188] The above has described various aspects of the present application with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each block in the flowchart and / or block diagram, and the combination of blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing devices to generate a machine, such that these instructions executed by the processor of the computer or other programmable data processing devices enable the implementation of the functions / actions specified in one or more blocks of the flowchart and / or block diagram. 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 should also be understood that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can also be implemented by dedicated hardware that performs the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0189] Those skilled in the art should understand that the above embodiments are all exemplary rather than restrictive. Different technical features appearing in different embodiments can be combined to achieve beneficial effects. Those skilled in the art should be able to understand and implement other variant embodiments of the disclosed embodiments based on the study of the drawings, the specification, and the claims. In the claims, the term "comprising" does not exclude other devices or steps; the quantifier "one" does not exclude a plurality; the terms "first" and "second" are used to label names rather than to indicate any specific order. Any reference signs in the claims should not be construed as limiting the scope of protection. The functions of multiple parts in the claims can be implemented by a single hardware or software module. The fact that certain technical features appear in different dependent claims does not mean that these technical features cannot be combined to achieve beneficial effects.

Claims

1. A verification method for real stores, characterized in that, Including: Obtain the store information of the first store as input, where the store information includes the store name and the store address; Based on the store name of the first store and the store address of the first store, search for the first target store name in the first target area in the electronic map, where the first target store name is the store name in the first target area whose similarity to the store name of the first store meets the first preset condition; Based on the store address of the first store, the preset real store address database, and the store name of the first store, search for the second target store name in the second target area in the electronic map, where the second target store name is the store name in the second target area whose similarity to the store name of the first store meets the second preset condition, and the real store address database includes the store addresses of real stores; In the case where the first target store name or the second target store name exists, determine that the first store is a real store; The step of searching for the second target store name in the second target area in the electronic map based on the store address of the first store, the preset real store address database, and the store name of the first store includes: Search in the preset real store address database for the first store address with the highest similarity to the store address of the first store; Use the electronic map to convert the first store address into a second geographical location; Calculate the similarity between the store name of the first store and the store names of each store in the third area where the second geographical location is located in the electronic map, and the second target area includes the third area; Search for the second target store name according to the similarity between the store name of the first store and the store names of each store in the third area and the second preset condition.

2. The method according to claim 1, characterized in that, The step of searching for the first target store name in the first target area in the electronic map based on the store name of the first store and the store address of the first store includes: Use the electronic map to convert the store address of the first store into a first geographical location; Calculate the similarity between the store name of the first store and the store names of each store in the first area where the first geographical location is located in the electronic map, and the first target area includes the first area; Search for the first target store name according to the similarity between the store name of the first store and the store names of each store in the first area and the first preset condition.

3. The method according to claim 2, characterized in that, The step of searching for the first target store name in the first target area in the electronic map based on the store name of the first store further includes: In the case where the first target store name is not found in the first area, calculate the similarity between the store name of the first store and the store names of each store in the second area surrounding the first area in the electronic map, and the first target area further includes the second area; Search for the first target store name according to the similarity between the store name of the first store and the store names of each store in the second area and the first preset condition.

4. The method according to claim 3, characterized in that, The first area is centered on the first geographical location; The inner boundary of the second region coincides with or is adjacent to the outer boundary of the first region, and the distance from the outer boundary of the first region to the first geographical location is less than the distance from the outer boundary of the second region to the first geographical location.

5. The method according to claim 1, characterized in that, The step of searching for the second target store name in the second target region on the electronic map based on the store address of the first store, the preset real store address database, and the store name of the first store further includes: In the case where the second target store name is not found in the third region, calculate the similarity between the store name of the first store and the store names of each store in the fourth region surrounding the third region, and the second target region further includes the fourth region; Search for the second target store name according to the similarity between the store name of the first store and the store names of each store in the fourth region and the second preset condition.

6. The method according to claim 5, characterized in that, The third region is centered on the second geographical location; The inner boundary of the fourth region coincides with or is adjacent to the outer boundary of the third region, and the distance from the outer boundary of the third region to the second geographical location is less than the distance from the outer boundary of the fourth region to the second geographical location.

7. The method according to claim 1, characterized in that, The first preset condition includes: having the highest similarity with the store name of the first store in the first target region and being greater than the real similarity threshold; The second preset condition includes: having the highest similarity with the store name of the first store in the second target region and being greater than the real similarity threshold.

8. The method according to claim 1, characterized in that, It further includes: In the case where both the first target store name and the second target store name do not exist, calculate the similarity between the store name of the first store and the store names of each store on the electronic map; Obtain the candidate stores on the electronic map, where the candidate stores include stores whose store names have a similarity higher than the real similarity threshold with the store name of the first store; Use the electronic map to obtain the geographical locations of the candidate stores; In the case where the distance between the geographical location of the candidate store and the first geographical location is less than or equal to the preset distance threshold, determine that the first store is a real store, and the first geographical location is the geographical location obtained by converting the store address of the first store using the electronic map; In the case where the distance between the geographical location of the candidate store and the first geographical location is greater than the preset distance threshold, determine that the first store is a non-real store.

9. The method according to any one of claims 1 to 8, characterized in that, The similarity is calculated based on the attribute similarity, The attribute similarity includes semantic similarity, feature structure similarity, and word order similarity.

10. The method according to claim 9, characterized in that, The similarity is the product of the a-th power of each of the attribute similarities, where a i is the weight parameter of the i-th attribute similarity. i ​ 11. The method according to claim 9, characterized in that, The semantic similarity is calculated according to the feature vectors obtained by converting the first object and the feature vectors obtained by converting the second object; The feature structure similarity is calculated according to the feature vectors obtained by converting the first object and the feature vectors obtained by converting the second object; The word order similarity is calculated according to the feature matrices obtained by converting the first object and the feature matrices obtained by converting the second object; Wherein, when the first object includes the store name of the first store, the second object includes the store name; when the first object includes the store address of the first store, the second object includes the store address.

12. The method according to claim 11, characterized in that, It further includes: Based on the feature vector obtained by converting the first object and the feature vector obtained by converting the second object, calculate the first norm of the feature vector obtained by converting the first object and the second norm of the feature vector obtained by converting the second object; According to the product of the first norm and the second norm, the square value of the first norm, the square value of the second norm, and a preset first smoothing coefficient, calculate the semantic similarity between the first object and the second object.

13. The method according to claim 11, characterized in that, It further includes: Based on the feature vector obtained by converting the first object and the feature vector obtained by converting the second object, calculate the first standard deviation of the feature vector obtained by converting the first object, the second standard deviation of the feature vector obtained by converting the second object, and the covariance between the feature vector obtained by converting the first object and the feature vector obtained by converting the second object; According to the product of the first standard deviation and the second standard deviation, the covariance, and a preset second smoothing coefficient, calculate the feature structure similarity between the first object and the second object.

14. The method according to claim 11, characterized in that, It further includes: Obtain a first word vector formed by column elements from the feature matrix obtained by converting the first object; Obtain a second word vector formed by column elements from the feature matrix obtained by converting the second object; According to the first word vector and the second word vector, calculate the word vector similarity between consecutive x first word vectors and consecutive x second word vectors, where x = 1, 2, 3,...; According to multiple word vector similarities, the text length of the first object, the text length of the second object, and the penalty term of the text lengths of the first object and the second object, calculate the word order similarity.

15. A verification device for a real store, characterized in that, It includes: An acquisition module for acquiring the store information of the input first store, where the store information includes the store name and the store address; A first search module for searching for a first target store name in a first target area in an electronic map based on the store name of the first store and the store address of the first store, where the first target store name is the store name in the first target area whose similarity to the store name of the first store meets a first preset condition; A second search module for searching for a second target store name in a second target area in an electronic map based on the store address of the first store, a preset real store address database, and the store name of the first store, where the second target store name is the store name in the second target area whose similarity to the store name of the first store meets a second preset condition, and the real store address database includes the store addresses of real stores; A determination module for determining that the first store is a real store when the first target store name or the second target store name exists. The second search module can be used to: search, from a preset real store address database, for a first store address that has the highest similarity to the store address of the first store; and use the electronic map to convert the first store address into a second geographical location; calculate the similarity between the store name of the first store and the store names of each store in a third area where the second geographical location is located in the electronic map, where the second target area includes the third area; and search for the second target store name according to the similarity between the store name of the first store and the store names of each store in the third area and the second preset condition.

16. A verification device for a real store, characterized in that, Comprising: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, the verification method of a real store as described in any one of claims 1 to 14 is implemented.

17. A computer-readable storage medium, characterized in that, Computer program instructions are stored on the computer-readable storage medium, and when the computer program instructions are executed by the processor, the verification method of a real store as described in any one of claims 1 to 14 is implemented.

Images