Address matching method and device based on repetition weight

Abstract

The present invention discloses a method and device for address matching based on repetition weights. The method comprises: vectorizing an address to be matched to obtain a corresponding address vector; calculating similarity between the address vector and candidate address vectors corresponding to multiple candidate addresses in a preset address database to obtain a weighted vector similarity between the address to be matched and any of the candidate addresses; the weighted vector similarity comprises the product of the similarity between the address vector and the candidate address vector and a repetition weight; the repetition weight is inversely proportional to the number of different physical objects that the same data appearance of the corresponding address vector or candidate address vector may simultaneously point to; and screening the target address corresponding to the address to be matched from the multiple candidate addresses based on the weighted vector similarity. It can be seen that the present invention can effectively improve the accuracy and efficiency of address matching.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and device for address matching based on repetition weight. Background Art

[0002] With the advancement of data processing algorithms and equipment performance, more and more transportation and transport companies are adopting data processing technologies to process address data. Address matching is a key component of this approach. Accurately matching ambiguous or erroneous addresses can effectively improve the efficiency and profitability of business operations. However, existing address matching technologies generally only calculate similarity between the data representing address characters, without considering the impact of repetitions between different addresses on similarity. Consequently, their matching accuracy is relatively low. Therefore, existing address matching methods based on repetition weighting have drawbacks that urgently need to be addressed. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide an address matching method and device based on repetition weight, which can effectively improve the accuracy and efficiency of address matching.

[0004] In order to solve the above technical problems, the first aspect of the present invention discloses an address matching method based on repetition weight, the method comprising:

[0005] Perform vector processing on the address to be matched to obtain the corresponding address vector;

[0006] Calculating similarity between the address vector and candidate address vectors corresponding to multiple candidate addresses in a preset address database to obtain a weighted vector similarity between the address to be matched and any of the candidate addresses; the weighted vector similarity comprises the product of the similarity between the address vector and the candidate address vector and a repetition weight; the repetition weight is inversely proportional to the number of different physical objects that may be simultaneously pointed to by the same data appearance of the corresponding address vector or the candidate address vector;

[0007] According to the weight vector similarity, a target address corresponding to the address to be matched is screened out from the multiple candidate addresses.

[0008] As an optional embodiment, in the first aspect of the present invention, the repetition weight includes a hierarchical repetition weight and / or a scene repetition weight; the hierarchical repetition weight is inversely proportional to the number of objects of different entity objects that the same data appearance of the address hierarchy to which the address vector or part or all of the candidate address vectors belongs may point to at the same time; the scene repetition weight is inversely proportional to the number of objects of different entity objects that the same data appearance of the scene type to which the address to be matched or part or all of the candidate addresses belongs may point to at the same time; the data appearance includes at least one of a data name, a data vector, a data identifier, and a data visualization pattern.

[0009] As an optional embodiment, in the first aspect of the present invention, the address vector includes a plurality of address fragment vectors corresponding to a plurality of address levels corresponding to the address to be matched; the candidate address vector includes a plurality of candidate address fragment vectors corresponding to a plurality of address levels corresponding to the candidate address;

[0010] Furthermore, the vectorization processing of the address to be matched to obtain the corresponding address vector includes:

[0011] Split the address to be matched to obtain address fragments to be matched corresponding to multiple address levels;

[0012] Vectorization is performed on the address fragment to be matched to obtain a corresponding address fragment vector.

[0013] As an optional embodiment, in the first aspect of the present invention, the calculating the similarity between the address vector and candidate address vectors corresponding to multiple candidate addresses in a preset address database to obtain a weighted vector similarity between the address to be matched and any of the candidate addresses includes:

[0014] For any candidate address among a plurality of candidate addresses in a preset address database, obtaining a candidate address fragment vector of candidate address fragments of a plurality of address levels corresponding to the candidate address;

[0015] Calculate the weighted vector similarity between any of the address fragment vectors corresponding to the address to be matched and the candidate address fragment vectors at the same address level; the weighted similarity is the product of the similarity between the address fragment vector and the candidate address fragment vector at the same level and the repetition weight.

[0016] The sum of the weighted vector similarities corresponding to at least two address segment vectors corresponding to the address to be matched is calculated to obtain the weighted vector similarity between the address to be matched and the candidate address.

[0017] As an optional embodiment, in the first aspect of the present invention, the similarity includes cosine distance and / or Euclidean distance; and / or, the hierarchical repetition weight is inversely proportional to the hierarchical refinement of the address level to which the address vector or part or all of the candidate address vectors belong in all address levels; and / or, the scene type includes a randomly named scene and a non-randomly named scene; wherein, the scene repetition weight corresponding to the randomly named scene is lower than the scene repetition weight corresponding to the non-randomly named scene.

[0018] As an optional implementation, in the first aspect of the present invention, the level duplication weight corresponding to any of the address levels can be calculated by the following steps:

[0019] For any of the address levels, obtaining information of multiple address fragments corresponding to the address level;

[0020] According to the physical address object corresponding to each of the address fragment information, a plurality of repeated fragment sets corresponding to the address level are screened out; each of the repeated fragment sets includes a plurality of address fragment information having the same data appearance but corresponding to different physical address objects;

[0021] Determining the level repetition weight corresponding to the address level according to the amount of the address segment information included in all the repeated segment sets;

[0022] And, determining the level repetition weight corresponding to the address level according to the number of the address segment information included in all the repeated segment sets includes:

[0023] Calculating a statistical value of the amount of the address fragment information included in all the repeated fragment sets; the statistical value includes at least one of a sum value, an average value, and a weighted average value;

[0024] The level repetition weight corresponding to the address level is determined according to the statistical value; the level repetition weight is inversely proportional to the statistical value.

[0025] As an optional implementation, in the first aspect of the present invention, screening out the target address corresponding to the to-be-matched address from the multiple candidate addresses based on the weight vector similarity includes:

[0026] Arrange the plurality of candidate addresses from largest to smallest according to the similarity of the weight vectors to obtain an address sequence;

[0027] Determine the first preset number of candidate addresses in the address sequence as target addresses corresponding to the address to be matched;

[0028] and / or,

[0029] At least one candidate address whose weight vector similarity is greater than a preset similarity threshold is screened out from the multiple candidate addresses and determined as the target address corresponding to the address to be matched.

[0030] A second aspect of the present invention discloses an address matching device based on repetition weight, the device comprising:

[0031] An address processing module is used to perform vector processing on the address to be matched to obtain the corresponding address vector;

[0032] a similarity calculation module, configured to perform similarity calculation on the address vector and candidate address vectors corresponding to a plurality of candidate addresses in a preset address database, to obtain a weighted vector similarity between the address to be matched and any of the candidate addresses; the weighted vector similarity comprising the product of the similarity between the address vector and the candidate address vector and a repetition weight; the repetition weight being inversely proportional to the number of different physical objects that the same data appearance of the corresponding address vector or candidate address vector may simultaneously point to;

[0033] The address screening module is used to screen out a target address corresponding to the to-be-matched address from the multiple candidate addresses according to the weight vector similarity.

[0034] As an optional embodiment, in the second aspect of the present invention, the repetition weight includes a hierarchical repetition weight and / or a scene repetition weight; the hierarchical repetition weight is inversely proportional to the number of objects of different physical objects that the same data appearance of the address hierarchy to which the address vector or part or all of the candidate address vectors belongs may point to at the same time; the scene repetition weight is inversely proportional to the number of objects of different physical objects that the same data appearance of the scene type to which the address to be matched or part or all of the candidate addresses belongs may point to at the same time; the data appearance includes at least one of a data name, a data vector, a data identifier, and a data visualization pattern.

[0035] As an optional embodiment, in the second aspect of the present invention, the address vector includes a plurality of address fragment vectors corresponding to a plurality of address levels corresponding to the address to be matched; the candidate address vector includes a plurality of candidate address fragment vectors corresponding to a plurality of address levels corresponding to the candidate address;

[0036] Furthermore, the address processing module performs vectorization processing on the address to be matched to obtain a specific method of a corresponding address vector, including:

[0037] Split the address to be matched to obtain address fragments to be matched corresponding to multiple address levels;

[0038] Vectorization is performed on the address fragment to be matched to obtain a corresponding address fragment vector.

[0039] As an optional embodiment, in the second aspect of the present invention, the similarity calculation module performs similarity calculation on the address vector with candidate address vectors corresponding to multiple candidate addresses in a preset address database to obtain a specific method of weight vector similarity between the address to be matched and any of the candidate addresses, including:

[0040] For any candidate address among a plurality of candidate addresses in a preset address database, obtaining a candidate address fragment vector of candidate address fragments of a plurality of address levels corresponding to the candidate address;

[0041] Calculate the weighted vector similarity between any of the address fragment vectors corresponding to the address to be matched and the candidate address fragment vectors at the same address level; the weighted similarity is the product of the similarity between the address fragment vector and the candidate address fragment vector at the same level and the repetition weight.

[0042] The sum of the weighted vector similarities corresponding to at least two address segment vectors corresponding to the address to be matched is calculated to obtain the weighted vector similarity between the address to be matched and the candidate address.

[0043] As an optional embodiment, in the second aspect of the present invention, the similarity includes cosine distance and / or Euclidean distance; and / or, the hierarchical repetition weight is inversely proportional to the hierarchical refinement of the address level to which the address vector or part or all of the candidate address vectors belong in all address levels; and / or, the scene type includes a randomly named scene and a non-randomly named scene; wherein, the scene repetition weight corresponding to the randomly named scene is lower than the scene repetition weight corresponding to the non-randomly named scene.

[0044] As an optional embodiment, in the second aspect of the present invention, the apparatus further includes a weight calculation module configured to perform the following steps to calculate the level duplication weight corresponding to any of the address levels:

[0045] For any of the address levels, obtaining information of multiple address fragments corresponding to the address level;

[0046] According to the physical address object corresponding to each of the address fragment information, a plurality of repeated fragment sets corresponding to the address level are screened out; each of the repeated fragment sets includes a plurality of address fragment information having the same data appearance but corresponding to different physical address objects;

[0047] Determining the level repetition weight corresponding to the address level according to the amount of the address segment information included in all the repeated segment sets;

[0048] Furthermore, the weight calculation module determines the level repetition weight corresponding to the address level according to the number of address segment information included in all the repeated segment sets, including:

[0049] Calculating a statistical value of the amount of the address fragment information included in all the repeated fragment sets; the statistical value includes at least one of a sum value, an average value, and a weighted average value;

[0050] The level repetition weight corresponding to the address level is determined according to the statistical value; the level repetition weight is inversely proportional to the statistical value.

[0051] As an optional embodiment, in the second aspect of the present invention, the specific manner in which the address screening module screens out the target address corresponding to the to-be-matched address from the multiple candidate addresses based on the weight vector similarity includes:

[0052] Arrange the plurality of candidate addresses from largest to smallest according to the similarity of the weight vectors to obtain an address sequence;

[0053] Determine the first preset number of candidate addresses in the address sequence as target addresses corresponding to the address to be matched;

[0054] and / or,

[0055] At least one candidate address whose weight vector similarity is greater than a preset similarity threshold is screened out from the multiple candidate addresses and determined as the target address corresponding to the address to be matched.

[0056] A third aspect of the present invention discloses another device for address matching based on repetition weight, the device comprising:

[0057] a memory storing executable program code;

[0058] a processor coupled to the memory;

[0059] The processor calls the executable program code stored in the memory to execute part or all of the steps in the address matching method based on repetition weight disclosed in the first aspect of the present invention.

[0060] The fourth aspect of the present invention discloses a computer-readable storage medium storing a computer program for electronic data exchange, wherein the computer program enables a computer to execute part or all of the steps in the address matching method based on repetition weight disclosed in the first aspect of the present invention.

[0061] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0062] In an embodiment of the present invention, a method and device for address matching based on repetition weight are disclosed. The method includes: vectorizing an address to be matched to obtain a corresponding address vector; calculating the similarity between the address vector and candidate address vectors corresponding to multiple candidate addresses in a preset address database to obtain a weight vector similarity between the address to be matched and any of the candidate addresses; the weight vector similarity includes the product of the similarity between the address vector and the candidate address vector and the repetition weight; the repetition weight is inversely proportional to the number of different entity objects that the same data appearance of the corresponding address vector or the candidate address vector may simultaneously point to; based on the weight vector similarity, the target address corresponding to the address to be matched is screened out from the multiple candidate addresses. It can be seen that the embodiment of the present invention can fully combine the repetition weight to calculate the similarity between the address to be matched and multiple candidate addresses, so as to more accurately determine the similarity between different candidate addresses and the address to be matched, thereby effectively improving the accuracy and efficiency of address matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0064] Figure 1 This is a flow chart of an address matching method based on repetition weight disclosed in an embodiment of the present invention.

[0065] Figure 2 It is a structural diagram of an address matching device based on repetition weight disclosed in an embodiment of the present invention.

[0066] Figure 3 It is a structural diagram of another address matching device based on repetition weight disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0067] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0068] The terms "second," "second," and the like in the specification and claims of the present invention and the accompanying drawings are used to distinguish between different objects, not to describe a specific order. Furthermore, the terms "including," "having," and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product, or device comprising a series of steps or elements is not limited to the listed steps or elements but may optionally include steps or elements not listed, or may optionally include other steps or elements inherent to such process, method, product, or device.

[0069] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0070] The present invention discloses a method and apparatus for address matching based on repetition weights. These methods utilize repetition weights to calculate the similarity between a target address and multiple candidate addresses, enabling a more accurate determination of the degree of similarity between different candidate addresses and the target address, thereby effectively improving the accuracy and efficiency of address matching. These methods are described in detail below.

[0071] Example 1

[0072] See also Figure 1 , Figure 1 This is a flow chart of an address matching method based on repetition weight disclosed in an embodiment of the present invention. Figure 1 The described address matching method based on repetition weight is applied to an address data processing chip, a processing terminal or a processing server (wherein the processing server may be a local server or a cloud server). Figure 1 As shown, the address matching method based on duplication weight may include the following operations:

[0073] 101. Perform vectorization processing on the address to be matched to obtain a corresponding address vector.

[0074] Optionally, the address to be matched can be input by the user or determined by the system based on preset rules. It is generally an address that has not yet been mapped to a physical address object and needs to be matched to determine its corresponding physical address object. Optionally, the address to be matched can include information at multiple address levels.

[0075] Optionally, the vectorization processing described in the present invention can be performed using a corresponding text or character vectorization algorithm model. For example, some pre-trained word vector models can be used to perform vectorization processing of addresses or address fragments, or some trained text prediction-related neural network model feature extractors can be used to perform vectorization processing.

[0076] 102. Perform similarity calculation on the address vector and candidate address vectors corresponding to multiple candidate addresses in a preset address database to obtain a weight vector similarity between the address to be matched and any candidate address.

[0077] Specifically, the weight vector similarity includes the product of the similarity between the address vector and the candidate address vector and the repetition weight, wherein the repetition weight is inversely proportional to the number of different entity objects that the same data appearance of the corresponding address vector or candidate address vector may point to at the same time.

[0078] By setting the duplication weight as described above, the degree of possible duplication corresponding to the address vector or candidate address vector can be effectively characterized. For example, the address corresponding to the address vector is likely to correspond to multiple different addresses, which means that it itself has a high degree of duplication. Therefore, its weight in calculating the overall similarity should be reduced. Therefore, setting the duplication weight inversely proportional to the degree of duplication can effectively improve the accuracy of the final calculated similarity.

[0079] Optionally, the repetition weight may include a level repetition weight and / or a scene repetition weight, for example, it may include the product of the level repetition weight and the scene repetition weight.

[0080] The hierarchical duplication weight is inversely proportional to the number of different physical objects that the same data appearance of the address hierarchy to which the address vector or some or all of the candidate address vectors belong may simultaneously point. The address hierarchy can be a manually or pre-defined hierarchy derived from different levels of address division. For example, it can be an administrative hierarchy, such as a city, county, or district, or a more detailed living area hierarchy, such as a house, bedroom, or bathroom, but this invention is not limited thereto.

[0081] Optionally, the hierarchical duplication weight can be inversely proportional to the hierarchical refinement of the address level to which the address vector or part or all of the vectors in the candidate address vector belong in all address levels. This is because the address level with a higher degree of hierarchical refinement is more likely to have duplicate names, so as to refer to more address objects. For example, there are fewer duplicate entities for provinces, cities, districts, and roads, which play a great role in determining the address, while the building number and room number have a high degree of duplicate names and have little effect on determining the address.

[0082] The scene repetition weight is inversely proportional to the number of different entity objects that may be simultaneously pointed to by the same data appearance of the scene type to which the address to be matched or part or all of the candidate addresses belong. Optionally, the scene type is used to indicate the scene function type of the address to be matched or part or all of the candidate addresses. It can be a scene in different dimensions, for example, it can be a community, a school, a hospital, or a restaurant. The degree of duplication of address names corresponding to different scenes may also vary, so this indicator can be set to characterize it.

[0083] Optionally, the scene types include arbitrarily named scenes and non-arbitrarily named scenes, wherein the scene repetition weight corresponding to the arbitrarily named scene is lower than the scene repetition weight corresponding to the non-arbitrarily named scene. Optionally, the arbitrarily named scene is used to indicate a scene address with a higher degree of freedom in naming, such as a self-operated restaurant or a private enterprise, so the possibility of having the same name is greater, and the scene repetition weight should be lower. Optionally, the non-arbitrarily named scene is used to indicate a scene address with a lower degree of freedom in naming, such as an address such as a community, school or hospital, so the possibility of having the same name is smaller, and the scene repetition weight should be higher.

[0084] Optionally, the data appearance of the present invention includes at least one of a data name, a data vector, a data identifier, and a data visualization pattern. Preferably, it is a data name or a data identifier.

[0085] 103. According to the weight vector similarity, a target address corresponding to the address to be matched is selected from multiple candidate addresses.

[0086] Optionally, the physical address object referred to by the target address can be determined as the physical address object to be referred to by the address to be matched, so as to determine the final practical meaning of the address to be matched, and also facilitate the subsequent execution of a series of business operations or data processing operations related to the address to be matched based on the physical address object corresponding to the address to be matched.

[0087] It can be seen that the above embodiments of the invention can fully combine the repetition weight to calculate the similarity between the address to be matched and multiple candidate addresses, so as to more accurately determine the similarity between different candidate addresses and the address to be matched, thereby effectively improving the accuracy and efficiency of address matching.

[0088] As an optional implementation, the address vector includes multiple address fragment vectors corresponding to multiple address levels corresponding to the to-be-matched address; the candidate address vector includes multiple candidate address fragment vectors corresponding to multiple address levels corresponding to the candidate address.

[0089] With the above settings, when similarity is subsequently calculated, similarity calculations can be performed on the address fragment vectors and candidate address fragment vectors at the same level and then statistics can be performed to obtain more accurate similarity.

[0090] Optionally, in step 101, vectorizing the address to be matched to obtain a corresponding address vector includes:

[0091] Split the address to be matched to obtain address fragments to be matched corresponding to multiple address levels;

[0092] Perform vectorization processing on the address fragment to be matched to obtain the corresponding address fragment vector.

[0093] Optionally, the address to be matched can be split by using an address semantic parsing algorithm model, for example, a pre-trained address segmentation neural network model or other algorithm model is used to split the address to be matched to obtain address fragments to be matched corresponding to multiple address levels.

[0094] It can be seen that through this optional implementation method, the address to be matched can be split to obtain address fragments to be matched corresponding to multiple address levels, and vectorized processing can be performed to obtain corresponding address fragment vectors, so that the similarity of address vectors at different levels can be calculated subsequently, which helps to calculate more accurate similarity, thereby effectively improving the accuracy and efficiency of address matching.

[0095] As an optional implementation, in step 102, similarity calculation is performed between the address vector and candidate address vectors corresponding to multiple candidate addresses in a preset address database to obtain a weighted vector similarity between the address to be matched and any candidate address, including:

[0096] For any candidate address among the multiple candidate addresses in the preset address database, obtaining a candidate address fragment vector of the candidate address fragments of the multiple address levels corresponding to the candidate address;

[0097] Calculate the weighted vector similarity between any address fragment vector corresponding to the address to be matched and the candidate address fragment vector at the same address level; the weighted similarity is the product of the similarity between the address fragment vector and the candidate address fragment vector at the same level and the repetition weight.

[0098] The sum of the weighted vector similarities corresponding to at least two address segment vectors corresponding to the address to be matched is calculated to obtain the weighted vector similarity between the address to be matched and the candidate address.

[0099] Optionally, the similarity described in the present invention may include cosine distance and / or Euclidean distance, which may be a weighted sum of one or both of them.

[0100] It can be seen that through this optional implementation, the sum of the weighted vector similarities corresponding to at least two address fragment vectors corresponding to the address to be matched can be calculated to obtain the weighted vector similarity between the address to be matched and the candidate address, so that a more accurate similarity can be calculated, thereby effectively improving the accuracy and efficiency of address matching.

[0101] As an optional implementation, the level duplication weight corresponding to any address level can be calculated by the following steps:

[0102] For any address level, obtain information about multiple address fragments corresponding to the address level;

[0103] According to the physical address object corresponding to each address fragment information, a plurality of repeated fragment sets corresponding to the address level are screened out; each repeated fragment set includes a plurality of address fragment information having the same data appearance but corresponding to different physical address objects;

[0104] The level repetition weight corresponding to the address level is determined according to the amount of address segment information included in all the repetition segment sets.

[0105] Optionally, the multiple address fragment information corresponding to the address level may be address fragment information associated with a specific address level and stored in a preset address database. The address fragment information may be address fragments of different data appearance types.

[0106] It can be seen that through this optional implementation, the hierarchical repetition weight corresponding to the address level can be determined according to the number of address fragment information included in all repeated fragment sets, so that a more accurate hierarchical repetition weight can be calculated, thereby effectively improving the accuracy and efficiency of address matching.

[0107] As an optional implementation, in the above step, determining the level repetition weight corresponding to the address level according to the number of address segment information included in all repeated segment sets includes:

[0108] Calculating a statistical value of the amount of address segment information included in all repeated segment sets;

[0109] According to the statistical value, the level duplication weight corresponding to the address level is determined.

[0110] Optionally, the statistical value may include at least one of a sum, an average, and a weighted average. Specifically, the level repetition weight should be inversely proportional to the statistical value. Optionally, the level repetition weight may be the inverse of the statistical value, or another inversely proportional mathematical relationship.

[0111] In a specific implementation, the average number of entities referred to by the same name in each field of the address database can be counted. The smaller the average number of duplicate names, the more limited the number of entities that can be referred to by a single name, and the more decisive it is for the accuracy of the address. For example, "Shanghai" is unique, while "Baoshan District" has duplicate names. Relatively speaking, the average value at the city level will be lower than the average value at the district level. The final hierarchical duplication weight can be taken as the inverse of the average value of duplicate names.

[0112] It can be seen that through this optional implementation, the hierarchical repetition weight corresponding to the address level can be determined based on the inverse mathematical relationship value of the statistical value of the number of address fragment information included in all repeated fragment sets, so that a more accurate hierarchical repetition weight can be calculated, which can effectively improve the accuracy and efficiency of address matching.

[0113] As an optional implementation, in step 103 above, selecting a target address corresponding to the to-be-matched address from a plurality of candidate addresses based on the weight vector similarity includes:

[0114] Arrange multiple candidate addresses from large to small according to the similarity of weight vectors to obtain an address sequence;

[0115] The first preset number of candidate addresses in the address sequence are determined as target addresses corresponding to the address to be matched.

[0116] It can be seen that through this optional implementation, the first preset number of candidate addresses with the highest weight vector similarity can be determined as the target addresses corresponding to the address to be matched, thereby effectively and accurately obtaining accurate address matching results and improving the accuracy and efficiency of address matching.

[0117] As an optional implementation, in step 103 above, selecting a target address corresponding to the to-be-matched address from a plurality of candidate addresses based on the weight vector similarity includes:

[0118] At least one candidate address whose weight vector similarity is greater than a preset similarity threshold is screened out from the multiple candidate addresses and determined as the target address corresponding to the address to be matched.

[0119] It can be seen that through this optional implementation, at least one candidate address whose weight vector similarity is greater than the preset similarity threshold can be screened out from multiple candidate addresses and determined as the target address corresponding to the address to be matched, thereby effectively and accurately obtaining an accurate address matching result and improving the accuracy and efficiency of address matching.

[0120] Example 2

[0121] See also Figure 2 , Figure 2 This is a schematic diagram of the structure of an address matching device based on repetition weight disclosed in an embodiment of the present invention. Figure 2 The described address matching device based on repetition weight is applied to an address data processing chip, a processing terminal or a processing server (wherein the processing server may be a local server or a cloud server). Figure 2 As shown, the address matching device based on repetition weight may include:

[0122] The address processing module 201 is used to perform vectorization processing on the address to be matched to obtain a corresponding address vector.

[0123] Optionally, the address to be matched can be input by the user or determined by the system based on preset rules. It is generally an address that has not yet been mapped to a physical address object and needs to be matched to determine its corresponding physical address object. Optionally, the address to be matched can include information at multiple address levels.

[0124] Optionally, the vectorization processing described in the present invention can be performed using a corresponding text or character vectorization algorithm model. For example, some pre-trained word vector models can be used to perform vectorization processing of addresses or address fragments, or some trained text prediction-related neural network model feature extractors can be used to perform vectorization processing.

[0125] The similarity calculation module 202 is configured to perform similarity calculation on the address vector and candidate address vectors corresponding to multiple candidate addresses in a preset address database to obtain a weight vector similarity between the address to be matched and any candidate address.

[0126] Specifically, the weight vector similarity includes the product of the similarity between the address vector and the candidate address vector and the repetition weight, wherein the repetition weight is inversely proportional to the number of different entity objects that the same data appearance of the corresponding address vector or candidate address vector may point to at the same time.

[0127] By setting the duplication weight as described above, the degree of possible duplication corresponding to the address vector or candidate address vector can be effectively characterized. For example, the address corresponding to the address vector is likely to correspond to multiple different addresses, which means that it itself has a high degree of duplication. Therefore, its weight in calculating the overall similarity should be reduced. Therefore, setting the duplication weight inversely proportional to the degree of duplication can effectively improve the accuracy of the final calculated similarity.

[0128] Optionally, the repetition weight may include a level repetition weight and / or a scene repetition weight, for example, it may include the product of the level repetition weight and the scene repetition weight.

[0129] The hierarchical duplication weight is inversely proportional to the number of different physical objects that the same data appearance of the address hierarchy to which the address vector or some or all of the candidate address vectors belong may simultaneously point. The address hierarchy can be a manually or pre-defined hierarchy derived from different levels of address division. For example, it can be an administrative hierarchy, such as a city, county, or district, or a more detailed living area hierarchy, such as a house, bedroom, or bathroom, but this invention is not limited thereto.

[0130] Optionally, the hierarchical duplication weight can be inversely proportional to the hierarchical refinement of the address level to which the address vector or part or all of the vectors in the candidate address vector belong in all address levels. This is because the address level with a higher degree of hierarchical refinement is more likely to have duplicate names, so as to refer to more address objects. For example, there are fewer duplicate entities for provinces, cities, districts, and roads, which play a great role in determining the address, while the building number and room number have a high degree of duplicate names and have little effect on determining the address.

[0131] The scene repetition weight is inversely proportional to the number of different entity objects that may be simultaneously pointed to by the same data appearance of the scene type to which the address to be matched or part or all of the candidate addresses belong. Optionally, the scene type is used to indicate the scene function type of the address to be matched or part or all of the candidate addresses. It can be a scene in different dimensions, for example, it can be a community, a school, a hospital, or a restaurant. The degree of duplication of address names corresponding to different scenes may also vary, so this indicator can be set to characterize it.

[0132] Optionally, the scene types include arbitrarily named scenes and non-arbitrarily named scenes, wherein the scene repetition weight corresponding to the arbitrarily named scene is lower than the scene repetition weight corresponding to the non-arbitrarily named scene. Optionally, the arbitrarily named scene is used to indicate a scene address with a higher degree of freedom in naming, such as a self-operated restaurant or a private enterprise, so the possibility of having the same name is greater, and the scene repetition weight should be lower. Optionally, the non-arbitrarily named scene is used to indicate a scene address with a lower degree of freedom in naming, such as an address such as a community, school or hospital, so the possibility of having the same name is smaller, and the scene repetition weight should be higher.

[0133] Optionally, the data appearance of the present invention includes at least one of a data name, a data vector, a data identifier, and a data visualization pattern. Preferably, it is a data name or a data identifier.

[0134] The address screening module 203 is used to screen out a target address corresponding to the to-be-matched address from a plurality of candidate addresses according to the weight vector similarity.

[0135] Optionally, the physical address object referred to by the target address can be determined as the physical address object to be referred to by the address to be matched, so as to determine the final practical meaning of the address to be matched, and also facilitate the subsequent execution of a series of business operations or data processing operations related to the address to be matched based on the physical address object corresponding to the address to be matched.

[0136] It can be seen that the above embodiments of the invention can fully combine the repetition weight to calculate the similarity between the address to be matched and multiple candidate addresses, so as to more accurately determine the similarity between different candidate addresses and the address to be matched, thereby effectively improving the accuracy and efficiency of address matching.

[0137] As an optional implementation, the address vector includes multiple address fragment vectors corresponding to multiple address levels corresponding to the to-be-matched address; the candidate address vector includes multiple candidate address fragment vectors corresponding to multiple address levels corresponding to the candidate address.

[0138] With the above settings, when similarity is subsequently calculated, similarity calculations can be performed on the address fragment vectors and candidate address fragment vectors at the same level and then statistics can be performed to obtain more accurate similarity.

[0139] Optionally, the address processing module 201 performs vectorization processing on the address to be matched to obtain a corresponding address vector in a specific manner including:

[0140] Split the address to be matched to obtain address fragments to be matched corresponding to multiple address levels;

[0141] Perform vectorization processing on the address fragment to be matched to obtain the corresponding address fragment vector.

[0142] Optionally, the address to be matched can be split by using an address semantic parsing algorithm model, for example, a pre-trained address segmentation neural network model or other algorithm model is used to split the address to be matched to obtain address fragments to be matched corresponding to multiple address levels.

[0143] It can be seen that through this optional implementation method, the address to be matched can be split to obtain address fragments to be matched corresponding to multiple address levels, and vectorized processing can be performed to obtain corresponding address fragment vectors, so that the similarity of address vectors at different levels can be calculated subsequently, which helps to calculate more accurate similarity, thereby effectively improving the accuracy and efficiency of address matching.

[0144] As an optional implementation, the similarity calculation module 202 calculates the similarity between the address vector and the candidate address vectors corresponding to multiple candidate addresses in a preset address database to obtain the specific method of weight vector similarity between the address to be matched and any candidate address, including:

[0145] For any candidate address among the multiple candidate addresses in the preset address database, obtaining a candidate address fragment vector of the candidate address fragments of the multiple address levels corresponding to the candidate address;

[0146] Calculate the weighted vector similarity between any address fragment vector corresponding to the address to be matched and the candidate address fragment vector at the same address level; the weighted similarity is the product of the similarity between the address fragment vector and the candidate address fragment vector at the same level and the repetition weight.

[0147] The sum of the weighted vector similarities corresponding to at least two address segment vectors corresponding to the address to be matched is calculated to obtain the weighted vector similarity between the address to be matched and the candidate address.

[0148] Optionally, the similarity described in the present invention may include cosine distance and / or Euclidean distance, which may be a weighted sum of one or both of them.

[0149] It can be seen that through this optional implementation, the sum of the weighted vector similarities corresponding to at least two address fragment vectors corresponding to the address to be matched can be calculated to obtain the weighted vector similarity between the address to be matched and the candidate address, so that a more accurate similarity can be calculated, thereby effectively improving the accuracy and efficiency of address matching.

[0150] As an optional embodiment, the device further includes a weight calculation module, configured to perform the following steps to calculate a level duplication weight corresponding to any address level:

[0151] For any address level, obtain information about multiple address fragments corresponding to the address level;

[0152] According to the physical address object corresponding to each address fragment information, a plurality of repeated fragment sets corresponding to the address level are screened out; each repeated fragment set includes a plurality of address fragment information having the same data appearance but corresponding to different physical address objects;

[0153] The level repetition weight corresponding to the address level is determined according to the amount of address segment information included in all the repetition segment sets.

[0154] Optionally, the multiple address fragment information corresponding to the address level may be address fragment information associated with a specific address level and stored in a preset address database. The address fragment information may be address fragments of different data appearance types.

[0155] It can be seen that through this optional implementation, the hierarchical repetition weight corresponding to the address level can be determined according to the number of address fragment information included in all repeated fragment sets, so that a more accurate hierarchical repetition weight can be calculated, thereby effectively improving the accuracy and efficiency of address matching.

[0156] As an optional implementation, the weight calculation module determines the level repetition weight corresponding to the address level according to the number of address segment information included in all repeated segment sets, including:

[0157] Calculating a statistical value of the amount of address segment information included in all repeated segment sets;

[0158] According to the statistical value, the level duplication weight corresponding to the address level is determined.

[0159] Optionally, the statistical value may include at least one of a sum, an average, and a weighted average. Specifically, the level repetition weight should be inversely proportional to the statistical value. Optionally, the level repetition weight may be the inverse of the statistical value, or another inversely proportional mathematical relationship.

[0160] It can be seen that through this optional implementation, the hierarchical repetition weight corresponding to the address level can be determined based on the inverse mathematical relationship value of the statistical value of the number of address fragment information included in all repeated fragment sets, so that a more accurate hierarchical repetition weight can be calculated, which can effectively improve the accuracy and efficiency of address matching.

[0161] As an optional implementation, the address screening module 203 may screen the target address corresponding to the to-be-matched address from multiple candidate addresses based on the weight vector similarity, including:

[0162] Arrange multiple candidate addresses from large to small according to the similarity of weight vectors to obtain an address sequence;

[0163] Determine the first preset number of candidate addresses in the address sequence as the target address corresponding to the address to be matched; arrange the multiple candidate addresses from large to small according to the similarity of the weight vector to obtain an address sequence;

[0164] The first preset number of candidate addresses in the address sequence are determined as target addresses corresponding to the address to be matched.

[0165] It can be seen that through this optional implementation, the first preset number of candidate addresses with the highest weight vector similarity can be determined as the target addresses corresponding to the address to be matched, thereby effectively and accurately obtaining accurate address matching results and improving the accuracy and efficiency of address matching.

[0166] As an optional implementation, the address screening module 203 may screen the target address corresponding to the to-be-matched address from multiple candidate addresses based on the weight vector similarity, including:

[0167] At least one candidate address whose weight vector similarity is greater than a preset similarity threshold is screened out from the multiple candidate addresses and determined as the target address corresponding to the address to be matched.

[0168] It can be seen that through this optional implementation, at least one candidate address whose weight vector similarity is greater than the preset similarity threshold can be screened out from multiple candidate addresses and determined as the target address corresponding to the address to be matched, thereby effectively and accurately obtaining an accurate address matching result and improving the accuracy and efficiency of address matching.

[0169] Example 3

[0170] See also Figure 3 , Figure 3 This is another address matching device based on repetition weight disclosed in an embodiment of the present invention. Figure 3 The described address matching device based on repetition weight is applied to an address data processing chip, a processing terminal or a processing server (wherein the processing server may be a local server or a cloud server). Figure 3 As shown, the address matching device based on repetition weight may include:

[0171] A memory 301 storing executable program code;

[0172] a processor 302 coupled to the memory 301;

[0173] The processor 302 calls the executable program code stored in the memory 301 to execute the steps of the address matching method based on repetition weight described in the first embodiment.

[0174] Example 4

[0175] An embodiment of the present invention discloses a computer-readable storage medium storing a computer program for electronic data exchange, wherein the computer program enables a computer to execute the steps of the address matching method based on repetition weight described in the first embodiment.

[0176] Example 5

[0177] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute the steps of the address matching method based on repetition weight described in the first embodiment.

[0178] The foregoing description of specific embodiments of the present disclosure is intended to illustrate a method for performing a multi-tasking process. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0179] The various embodiments in this specification are described in a progressive manner. Similar portions between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from the other embodiments. In particular, the device, apparatus, and non-volatile computer-readable storage medium embodiments are generally similar to the method embodiments, so their descriptions are relatively simplified. For relevant details, refer to the descriptions of the method embodiments.

[0180] The apparatus, device, non-volatile computer-readable storage medium and method provided in the embodiments of this specification correspond to each other. Therefore, the apparatus, device, and non-volatile computer storage medium also have similar beneficial technical effects as the corresponding method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the corresponding apparatus, device, and non-volatile computer storage medium will not be repeated here.

[0181] In the 1990s, technological improvements could be clearly distinguished as either hardware improvements (for example, improvements to circuit structures like diodes, transistors, and switches) or software improvements (improvements to process flows). However, with the advancement of technology, many process flow improvements today can now be considered direct improvements to hardware circuit structures. Designers almost always create the corresponding hardware circuit structure by programming the improved process flow into the hardware circuit. Therefore, it cannot be said that a process flow improvement cannot be implemented using a hardware module. For example, a programmable logic device (PLD), such as a field programmable gate array (FPGA), is an integrated circuit whose logical function is determined by user programming. Designers can "integrate" a digital system on a PLD through their own programming, eliminating the need for a chip manufacturer to design and manufacture a dedicated integrated circuit chip. Moreover, nowadays, instead of manually fabricating integrated circuit chips, this programming is mostly done using "logic compiler" software. This is similar to the software compiler used when developing programs. Before compilation, the original code must also be written in a specific programming language, called a hardware description language (HDL). There is not just one HDL, but many, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. Currently, the most commonly used are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art will also understand that by simply programming the method flow in one of these hardware description languages ​​and then programming it into an integrated circuit, a hardware circuit that implements the logic method flow can be easily obtained.

[0182] The controller can be implemented in any suitable manner. For example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also know that in addition to implementing the controller in a purely computer-readable program code format, the controller can be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, such a controller can be considered a hardware component, and the devices included therein for implementing various functions can also be considered as structures within the hardware component. Or even, the devices for implementing various functions can be considered as both software modules that implement the method and structures within the hardware component.

[0183] The systems, devices, modules, or units described in the above embodiments may be implemented by computer chips or entities, or by products having certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0184] For the convenience of description, the above devices are described as being divided into various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0185] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the embodiments of this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0186] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, 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, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0187] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0188] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0189] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0190] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0191] Computer-readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0192] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0193] This specification may be described in the general context of computer-executable instructions, such as program modules, executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media, including storage devices.

[0194] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0195] Finally, it should be noted that the address matching method and device based on repetition weight disclosed in the embodiment of the present invention are only preferred embodiments of the present invention, and are only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, it should be understood by those skilled in the art that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. An address matching method based on repetition weight, characterized in that: The method comprises: Perform vectorization processing on the address to be matched to obtain the corresponding address vector; Calculating similarity between the address vector and candidate address vectors corresponding to multiple candidate addresses in a preset address database to obtain a weighted vector similarity between the address to be matched and any of the candidate addresses; the weighted vector similarity comprises the product of the similarity between the address vector and the candidate address vector and a repetition weight; the repetition weight is inversely proportional to the number of different physical objects that may be simultaneously pointed to by the same data appearance of the corresponding address vector or the candidate address vector; Filtering a target address corresponding to the to-be-matched address from the multiple candidate addresses according to the weight vector similarity; The repetition weight includes a hierarchical repetition weight and a scene repetition weight; the hierarchical repetition weight is inversely proportional to the number of different entity objects that may be simultaneously pointed to by the same data appearance of the address hierarchy to which the address vector or part or all of the candidate address vectors belong; the scene repetition weight is inversely proportional to the number of different entity objects that may be simultaneously pointed to by the same data appearance of the scene type to which the to-be-matched address or part or all of the candidate addresses belong; the data appearance includes at least one of a data name, a data vector, a data identifier, and a data visualization pattern; The level duplication weight corresponding to any of the address levels is calculated by the following steps: For any of the address levels, obtaining information of multiple address fragments corresponding to the address level; According to the physical address object corresponding to each of the address fragment information, a plurality of repeated fragment sets corresponding to the address level are screened out; each of the repeated fragment sets includes a plurality of address fragment information having the same data appearance but corresponding to different physical address objects; Determining the level repetition weight corresponding to the address level according to the amount of the address segment information included in all the repeated segment sets; And, determining the level repetition weight corresponding to the address level according to the number of the address segment information included in all the repeated segment sets includes: Calculating a statistical value of the amount of the address fragment information included in all the repeated fragment sets; the statistical value includes at least one of a sum value, an average value, and a weighted average value; The level repetition weight corresponding to the address level is determined according to the statistical value; the level repetition weight is inversely proportional to the statistical value.

2. The address matching method based on repetition weight according to claim 1, characterized in that: The address vector includes a plurality of address fragment vectors corresponding to a plurality of address levels corresponding to the address to be matched; the candidate address vector includes a plurality of candidate address fragment vectors corresponding to a plurality of address levels corresponding to the candidate address; Furthermore, the vectorization processing of the address to be matched to obtain the corresponding address vector includes: Split the address to be matched to obtain address fragments to be matched corresponding to multiple address levels; Vectorization is performed on the address fragment to be matched to obtain a corresponding address fragment vector.

3. The address matching method based on repetition weight according to claim 2, characterized in that: The calculating the similarity between the address vector and candidate address vectors corresponding to a plurality of candidate addresses in a preset address database to obtain a weight vector similarity between the address to be matched and any of the candidate addresses includes: For any candidate address among a plurality of candidate addresses in a preset address database, obtaining a candidate address fragment vector of candidate address fragments of a plurality of address levels corresponding to the candidate address; Calculating a weighted vector similarity between any of the address fragment vectors corresponding to the address to be matched and the candidate address fragment vectors at the same address level; the weighted vector similarity is the product of the similarity between the address fragment vector and the candidate address fragment vector at the same level and the repetition weight; The sum of the weighted vector similarities corresponding to at least two address segment vectors corresponding to the address to be matched is calculated to obtain the weighted vector similarity between the address to be matched and the candidate address.

4. The address matching method based on repetition weight according to claim 1, characterized in that: The similarity includes cosine distance and / or Euclidean distance; and / or, the hierarchical repetition weight is inversely proportional to the hierarchical refinement of the address level to which the address vector or part or all of the candidate address vectors belong in all address levels; and / or, the scene type includes a randomly named scene and a non-randomly named scene; wherein, the scene repetition weight corresponding to the randomly named scene is lower than the scene repetition weight corresponding to the non-randomly named scene.

5. The address matching method based on repetition weight according to claim 1, characterized in that: The step of selecting a target address corresponding to the to-be-matched address from the plurality of candidate addresses according to the weight vector similarity includes: Arrange the plurality of candidate addresses from largest to smallest according to the similarity of the weight vectors to obtain an address sequence; Determine the first preset number of candidate addresses in the address sequence as target addresses corresponding to the address to be matched; and / or, At least one candidate address whose weight vector similarity is greater than a preset similarity threshold is screened out from the multiple candidate addresses and determined as the target address corresponding to the address to be matched.

6. An address matching device based on repetition weight, characterized in that: The device is used to execute the address matching method based on repetition weight according to any one of claims 1 to 5, and the device includes: An address processing module is used to perform vector processing on the address to be matched to obtain the corresponding address vector; a similarity calculation module, configured to perform similarity calculation on the address vector and candidate address vectors corresponding to a plurality of candidate addresses in a preset address database, to obtain a weighted vector similarity between the address to be matched and any of the candidate addresses; the weighted vector similarity comprising the product of the similarity between the address vector and the candidate address vector and a repetition weight; the repetition weight being inversely proportional to the number of different physical objects that the same data appearance of the corresponding address vector or candidate address vector may simultaneously point to; The address screening module is used to screen out a target address corresponding to the to-be-matched address from the multiple candidate addresses according to the weight vector similarity.

7. An address matching device based on repetition weight, characterized in that: The device comprises: a memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the address matching method based on repetition weight according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that: The computer program for electronic data exchange is stored therein, wherein the computer program enables a computer to execute the address matching method based on repetition weight as claimed in any one of claims 1 to 5.

Images