Address processing method, device, storage medium and electronic device

By matching the entity recognition model and the address data set, the problem of low accuracy in address recognition is solved, and the accurate identification and matching of address entities is achieved, and the accuracy of address recognition is improved.

CN114722824BActive Publication Date: 2024-11-26BEIJING JINTI TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210344873.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-31
Publication Date
2024-11-26
Estimated Expiration
2042-03-31

AI Technical Summary

Technical Problem

In the existing address recognition methods, the address matching accuracy is low, and it is easy to misidentify non-address entities as addresses or identify addresses as other types, resulting in inaccurate identification.

Method used

Through the entity recognition model, the query message is identified in entity category, the address entity category is located, and the address subset is matched according to the address data set, and the category correction is carried out for other categories of entities located behind the address entity to ensure the accurate identification of address entities.

Benefits of technology

Improve the accuracy of address recognition, reduce the errors in address recognition as other types, and enhance the accuracy of address matching.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114722824B_ABST
    Figure CN114722824B_ABST
Patent Text Reader

Abstract

The present invention relates to an address processing method, device, storage medium and electronic device, and relates to the field of data processing. The method comprises: obtaining a query message, performing entity recognition on information in the query message to obtain the category of each entity of the information, and in the case where there is a first entity of address category in the information, matching the address subset corresponding to the first entity according to the address data set, and in the case where the address subset corresponding to the first entity is matched, performing category correction on entities of other categories located after the first entity in the information according to the address subset. The address processing method, device, storage medium and electronic device provided by the present invention can solve the problem of low accuracy of address matching in address recognition.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an address processing method, device, storage medium and electronic device. Background Art

[0002] In company search scenarios, it is often necessary to identify the addresses in user search content in order to find relevant information more accurately and improve the user search experience. However, in current address recognition methods, keywords are usually used for matching. However, due to the large number of administrative districts at all levels and the variety of names, it is easy to misidentify non-address entities or misidentify address entities as other types, which easily leads to low address matching accuracy in address recognition. Summary of the invention

[0003] The object of the present invention is to provide an address processing method, device, storage medium and electronic device for solving the problem of low accuracy of address matching in address recognition.

[0004] According to a first aspect of an embodiment of the present invention, there is provided an address processing method, including:

[0005] Get query message;

[0006] Performing entity recognition on the information in the query message to obtain categories of each entity in the information;

[0007] In the case where a first entity of a category of address exists in the information, matching an address subset corresponding to the first entity according to the address data set;

[0008] In the case where the address subset corresponding to the first entity is matched, the categories of entities of other categories located after the first entity in the information are modified according to the address subset.

[0009] Optionally, performing entity recognition on the information in the query message to obtain categories of each entity of the information includes:

[0010] The information in the query message is subjected to entity recognition through an entity recognition model to obtain the category of each entity in the information output by the entity recognition model.

[0011] Optionally, the method further comprises:

[0012] When there are multiple consecutive entities of the address category, determine the last entity of the multiple consecutive entities as the first entity;

[0013] When there are multiple non-continuous entities of the address type, all of the multiple non-continuous entities are taken as the first entity.

[0014] Optionally, determining the last entity of the multiple consecutive entities as the first entity includes:

[0015] In the case that the last entity belongs to at least two upper administrative district levels, the last n entities of the multiple continuous entities are collectively regarded as the first entity.

[0016] Optionally, in the case where the address subset corresponding to the first entity is matched, modifying the category of entities of other categories located after the first entity in the information according to the address subset includes:

[0017] In the case where the address subset corresponding to the first entity is matched, searching the address subset corresponding to the first entity to see whether there is an address matching a second entity, where the second entity is an entity located after the first entity and adjacent to the first entity;

[0018] If there is an address matching the second entity in the address subset corresponding to the first entity, correct the type of the second entity to an address;

[0019] Querying the address data set whether a corresponding address subset exists for the second entity;

[0020] In the case that it is found that the second entity has a corresponding address subset, the second entity is taken as a new first entity, and the step of searching the address subset corresponding to the first entity to see whether there is an address matching the second entity is repeated.

[0021] Optionally, when there is no address matching the second entity in the address subset corresponding to the first entity, or when the address subset corresponding to the second entity cannot be found, the category modification ends.

[0022] Optionally, the method further comprises:

[0023] Determining a target address corresponding to the information according to the category of each entity after category correction;

[0024] Acquire at least one related enterprise information according to the target address;

[0025] The at least one enterprise information is outputted.

[0026] Optionally, determining the target address corresponding to the information according to the category of each entity after category modification includes:

[0027] After the category modification is performed, multiple entities of the address category among the entities are combined according to the administrative district level to generate the target address.

[0028] Optionally, after performing the category modification, combining entities of the category of addresses among the entities according to the administrative district level to generate the target address includes:

[0029] After performing the category correction, determining whether there is an address missing between any two entities among the multiple entities according to the administrative district levels to which the multiple entities belong;

[0030] When it is determined that there are missing addresses between any two entities, obtaining corresponding missing addresses from the address data set according to the multiple entities and supplementing the missing addresses between the two entities;

[0031] The target address is generated according to the multiple entities and the supplemented missing address.

[0032] Optionally, the address data set includes address subsets corresponding to multiple addresses belonging to various administrative levels, and the various administrative levels include province, city, county or district, township or town, and village.

[0033] Optionally, the address subset corresponding to each of the multiple addresses includes: multiple addresses and / or address abbreviations belonging to the next administrative district level of the address.

[0034] According to a second aspect of an embodiment of the present invention, there is provided an address processing device, the device comprising:

[0035] An acquisition module, used to acquire query messages;

[0036] An identification module, used to perform entity identification on the information in the query message to obtain the category of each entity of the information;

[0037] A matching module, configured to match, when a first entity of address category exists in the information, a subset of addresses corresponding to the first entity according to the address data set;

[0038] The processing module is used to modify the categories of entities of other categories located after the first entity in the information according to the address subset when matching the address subset corresponding to the first entity.

[0039] According to a third aspect of an embodiment of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the steps of the method described in any one of the first aspects are implemented.

[0040] According to a fourth aspect of an embodiment of the present invention, there is provided an electronic device, including:

[0041] a memory having a computer program stored thereon;

[0042] A processor is used to execute the computer program in the memory to implement the steps of any one of the methods described in the first aspect above.

[0043] Through the above technical solution, after obtaining the query message, the information in the query message is entity identified to obtain the categories of each entity in the information. When there is a first entity of address category in the information, the address subset corresponding to the first entity is matched according to the address data set, and when the address subset corresponding to the first entity is matched, the category of entities of other categories located after the first entity in the information is corrected according to the address subset. The address processing method, device, storage medium and electronic device provided by the present invention first use the entity recognition model to identify the entity category for the query information input by the user, preliminarily determine the entity type in the query information, locate the address entity category, and then match the entity after the address entity that is not identified as an address through the address data set to determine whether to correct the category to an address, which can solve the problem of identifying the address as other entity types, thereby improving the address matching accuracy in address recognition.

[0044] Other features and advantages of the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the present invention but do not constitute a limitation of the present invention. In the accompanying drawings:

[0046] Figure 1 is a flowchart of an address processing method according to an exemplary embodiment;

[0047] Figure 2 is a flowchart of another address processing method according to an exemplary embodiment;

[0048] Figure 3 is a flowchart of another address processing method according to an exemplary embodiment;

[0049] Figure 4 is a flowchart of another address processing method according to an exemplary embodiment;

[0050] Figure 5 is a block diagram of an address processing device according to an exemplary embodiment;

[0051] Figure 6 It is a block diagram of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION

[0052] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0053] In user search scenarios, it is often necessary to identify the addresses in the search content, such as identifying the addresses contained in the company names searched by users, so as to search for related companies more accurately, or to determine other related companies in the location based on the address information, so as to facilitate users to view relevant information and improve the search experience.

[0054] In the related art, a keyword matching method is used for address recognition. However, because there are a large number of provincial, municipal, county and township names, and many place names are strange, it is easy to mistakenly recognize non-address entities, such as "Baiyin City". Since "Baiyin" is also a very common industry vocabulary, when the keyword matching method is directly used for recognition, there is a certain probability that the matching result is wrong, and the address entity cannot be recognized, resulting in low address recognition accuracy. Therefore, in order to solve the above problem, the present invention provides an address processing method, which is described below.

[0055] Figure 1 is a flowchart of an address processing method according to an exemplary embodiment. Figure 1 As shown, the address processing method includes the following steps.

[0056] In step S11, a query message is obtained.

[0057] The query message may be any message input by a user, and may include multiple entity categories, such as address entities, industry entities, name entities, font size entities, etc., wherein there may be multiple address entities, and the multiple address entities may be continuous or non-continuous.

[0058] In step S12, entity recognition is performed on the information in the query message to obtain the category of each entity in the information.

[0059] It is understandable that since the query message may contain multiple entity types, the type of each entity in the query message can be identified first. For example, the query message can be subjected to entity recognition through a pre-trained entity recognition model to identify the category of each entity in the query message. The entity category may include address, industry, business name, name, and other possible categories. The division of entity categories can be determined based on actual conditions, and the present invention does not limit this.

[0060] In step S13, when there is a first entity of address type in the information, the address subset corresponding to the first entity is matched according to the address data set.

[0061] Among them, the address data set can be understood as an address dictionary, and the address data set may include address subsets corresponding to multiple addresses belonging to multiple administrative district levels, wherein the multiple administrative district levels may include: country, province, city, county or district, township or town, village. Each administrative district level corresponds to multiple addresses. For example, the provincial administrative district may include all provinces (or municipalities) in the country, and the name of each province can be used as a provincial address, so each province corresponds to an address subset, and the address subset corresponding to each province includes each city under the province, and the name of each city can be used as a city-level address, so each city corresponds to an address subset, and the address subset corresponding to each city includes each district or county of the city, and the same applies to the district level (or county level) and township level, and so on. For example, after determining the first entity in the query message, assuming that the first entity is a provincial address, the city-level address subset corresponding to the provincial address is matched according to the above-mentioned address data set.

[0062] Optionally, for each of the multiple addresses, the address subset corresponding to the address includes: multiple addresses and / or address abbreviations belonging to the next administrative district level of the address.

[0063] For example, the administrative level can be provincial level. Taking Shaanxi Province as an example, the address subset corresponding to the address can include the cities under the jurisdiction of Shaanxi Province, such as Xi'an, Xianyang, Weinan and Baoji. When the administrative level is city level, taking Xi'an as an example, the address subset corresponding to Xi'an can include the districts under the jurisdiction of Xi'an, such as Yanta District, Beilin District, Lianhu District and Chang'an District.

[0064] In step S14, when the address subset corresponding to the first entity is matched, the categories of entities of other categories located after the first entity in the information are modified according to the address subset.

[0065] For example, taking the provincial address in the query message in step S13 as an example, after matching the municipal address subset corresponding to the provincial address according to the address data set, the categories of entities of other categories in the query message are corrected according to the municipal address subset. It can be understood that based on the writing habits of addresses, they are generally input from high administrative district levels to low administrative district levels, such as A province B city C district, so the address entities input later in the order are usually more specific addresses, and more specific addresses can make address recognition more accurate. Therefore, after determining the first entity, the categories of entities of other categories located after the first entity can be corrected based on the queried address subset.

[0066] Figure 2 is a flowchart of another address processing method according to an exemplary embodiment. Figure 2 As shown, the entity recognition of the information in the query message in step S12 to obtain the category of each entity of the information may include the following steps:

[0067] Step S121, performing entity recognition on the information in the query message through an entity recognition model to obtain the category of each entity in the information output by the entity recognition model.

[0068] Among them, named entity recognition (NER) is a classic problem in natural language processing. The entity recognition model used in this embodiment can be any model with entity recognition function, such as a conditional random field (CRF) model. The entity recognition model can be selected according to actual needs, and the present invention does not limit it. For example, the query message is subjected to entity recognition through a pre-trained entity recognition model to identify the category of each entity in the query message. The entity category may include address, industry, font size, name, and other possible categories. The division of entity categories can be determined according to actual conditions.

[0069] For example, the query message content input by the user is "Beijing Haidian XXX Technology Company". By identifying the query message through the trained entity recognition model, the address entities "Beijing" and "Haidian", the industry entity "Technology Company" and the font size entity "XXX" can be obtained.

[0070] Step S122: when there are multiple consecutive entities of the address type, determine the last entity of the multiple consecutive entities as the first entity.

[0071] For example, since the query message input by the user may contain multiple address entities, and the multiple address entities may be continuous, and usually multiple continuous address entities correspond to multiple administrative regions, the last entity of the multiple continuous entities can be determined as the first entity, that is, the last address entity of the multiple continuous entities is used to query whether there is a corresponding address subset. For example, the information in the query message is "AA Culture Co., Ltd., Xi'an, Shaanxi Province, China". In the address entity category, "China", "Shaanxi Province", and "Xi'an City" belong to multiple continuous entities, "AA" is identified as a font size entity, and "Culture" is identified as an industry entity, so that "Xi'an City" is determined as the first entity, and the corresponding address subset is matched according to "Xi'an City" to correct the category of the entity after "Xi'an City".

[0072] Optionally, since there may be special cases where address names have the same name, when the last entity belongs to at least two higher-level administrative district levels, the last n entities in the multiple consecutive entities are collectively used as the first entity. Wherein, the last entity belongs to at least two higher-level administrative district levels, which can be understood as the at least two higher-level administrative district levels both containing an address with the same name as the last entity. For example, there is "Nanjing Road" in both Shanghai and Tianjin. Therefore, if the last entity is "Nanjing Road", there are two higher-level administrative district levels to which the entity belongs. In this case, it is necessary to introduce more address entities to jointly query the address subset. Therefore, the last n entities in the multiple consecutive entities can be collectively used as the first entity, where n is a positive integer, such as 2 or 3.

[0073] For example, the information in the query message includes "Nanjing Road, Huangpu District, Shanghai", then the multiple continuous entities of the address category are "Shanghai", "Huangpu District", and "Nanjing Road". Since the address entity "Nanjing Road" also exists in other cities besides Shanghai, such as "Nanjing Road, Heping District, Tianjin", it is necessary to use "Huangpu District" and "Nanjing Road" as the first entity to query the corresponding address subset, or to use the three address entities "Shanghai", "Huangpu District", and "Nanjing Road" as the first entity to query the corresponding address subset.

[0074] Step S123: when there are multiple non-continuous entities of the address type, all of the multiple non-continuous entities are regarded as the first entity.

[0075] It is understandable that when a user inputs an address, if they input multiple address entities, they are generally multiple consecutive address entities according to input habits. When the address entities are discontinuous, there may be a possibility of entity category recognition errors between the discontinuous entities. Therefore, the multiple discontinuous entities can be used as the first entity to match the corresponding address subset and perform category correction, so as to correct the address entities that may be incorrectly identified between the discontinuous entities.

[0076] For example, the information in the query message is "BB Technology Company, Baiyin District, Baiyin City, Gansu Province", and the results identified by the entity recognition model are assumed to be the address entity "Gansu Province", the precious metal entity "Baiyin", the address entity "Baiyin District", the font size entity "BB", and the industry entity "Technology Company". Among them, the entities "Gansu Province" and "Baiyin District" with the address category are non-continuous, so "Gansu Province" and "Baiyin District" can be used as the first entity, so that "Gansu Province" and "Baiyin District" can be used to match the corresponding address subsets respectively. After matching the address subset corresponding to "Gansu Province", Baiyin City can be found according to the names of the cities in Gansu Province contained in the address subset, so that the "Baiyin" after "Gansu Province" in the query message can be corrected to an address entity. After matching the address subset corresponding to "Baiyin District", the address subset is used to correct the category of the entity after "Baiyin District".

[0077] Figure 3 is a flowchart of another address processing method according to an exemplary embodiment. Figure 3 As shown, in step S14, when the address subset corresponding to the first entity is matched, the category of entities of other categories located after the first entity in the information is modified according to the address subset, which may include the following steps.

[0078] Step S141: when the address subset corresponding to the first entity is matched, searching the address subset corresponding to the first entity to see whether there is an address matching the second entity.

[0079] The second entity is an entity located behind the first entity and adjacent to the first entity.

[0080] For example, the information content in the query message is "Shaanxi History Museum, Yanta, Xi'an, Shaanxi Province". Entity recognition is performed through the entity recognition model, and the recognition results obtained can be the address entities "Shaanxi Province", "Xi'an City", the building entity "Yanta", and the building entity "Shaanxi History Museum". The last entity "Xi'an City" in the consecutive address entities "Shaanxi Province" and "Xi'an City" can be used as the first entity to search for the corresponding address subset in the address data set. The address subset matched by the first entity can include 11 districts and 2 counties under the jurisdiction of Xi'an Province: Xincheng District, Beilin District, Lianhu District, Yanta District, Baqiao District, Weiyang District, Yanliang District, Lintong District, Chang'an District, Gaoling District, Huyi District, Lantian County, and Zhouzhi County. Then, the address subset can be further searched for addresses that match the second entity "Yanta".

[0081] It is worth mentioning that in the address data set, the correspondence between the address at any administrative district level and its address subset can be in the form of a key-value pair. For the address at any administrative district level, the address or the address abbreviation can be used as the key, and the address subset can be used as the value. Therefore, after determining the first entity, the corresponding address subset can be queried using the first entity as the key. The elements in the address subset can include the address or address abbreviation of the sub-area under the first entity.

[0082] For example, taking Shaanxi Province as the first entity, the corresponding address subset can be expressed as: {"Shaanxi Province", "Shaanxi": ["Xi'an City", "Xi'an", "Xianyang City", "Xianyang", ...]}, taking Xi'an City as the first entity, the corresponding address subset can be expressed as {"Xi'an City", "Xi'an": ["Beilin District", "Beilin", "Yanta District", "Yanta", ...,]}.

[0083] In addition, if there is no address matching the second entity in the address subset corresponding to the first entity, step S145 is executed.

[0084] Step S142: if there is an address matching the second entity in the address subset corresponding to the first entity, the type of the second entity is corrected to an address.

[0085] For example, in the address subset corresponding to the first entity "Shaanxi Province" mentioned above, it can be determined that there is an address "Yanta District" that matches the second entity "Yanta", so the category of the second entity "Yanta" can be corrected from a building entity to an address entity.

[0086] Step S143: query the address data set to see whether there is a corresponding address subset for the second entity.

[0087] For example, after the second entity is corrected to an address entity, the address subset corresponding to the second entity "Yanta District" can be further searched in the address data set, and the address subset can include various streets under the jurisdiction of "Yanta District".

[0088] Step S144: when the address subset corresponding to the second entity is found, the second entity is taken as a new first entity, and the step of searching the address subset corresponding to the first entity to see whether there is an address matching the second entity is repeated.

[0089] For example, when there is a corresponding address subset for the above-mentioned second entity "Yanta District", "Yanta District" is taken as the new first entity, and the next entity "Shaanxi History Museum" located after the new first entity "Yanta District" in the query message is taken as the new second entity, and the step of searching whether there is an address matching the second entity in the address subset corresponding to the first entity described in S141 is repeated. That is, it can be understood as searching whether there is a street and "Shaanxi History Museum" in the address subset corresponding to "Yanta District". Since "Shaanxi History Museum" does not belong to a street address, the address subset corresponding to "Yanta District" cannot find an address matching the "Shaanxi History Museum", that is, for the new first entity and the new second entity, there is no address matching the second entity in the address subset corresponding to the first entity.

[0090] In addition, if the address subset corresponding to the second entity cannot be found, step S145 is executed.

[0091] Step S145, category correction ends.

[0092] Figure 4 is a flowchart of another address processing method according to an exemplary embodiment. Figure 4 As shown, after step S14, the method may further include the following steps:

[0093] In step S15, according to the categories of each entity after category modification, the target address corresponding to the information is determined. For example, the following steps may be included:

[0094] After the category correction is performed, multiple entities of the address category in each entity are combined according to the administrative district level to generate a target address.

[0095] The combination can be combined in order from high to low administrative district levels. For example, after category correction, the address entities can be "Beijing City", "Haidian District", and "Zhongguancun South Street". By combining these three entities of the address category in the administrative district hierarchy order of city-district-street, the target address "Zhongguancun South Street, Haidian District, Beijing City" can be generated.

[0096] In step S16, at least one relevant enterprise information is acquired according to the target address.

[0097] In step S17, at least one piece of enterprise information is output.

[0098] For example, a business information query is performed based on the target address "Zhongguancun South Street, Haidian District, Beijing" to obtain at least one business whose address contains the information "Zhongguancun South Street, Haidian District, Beijing", and the obtained relevant business information is output.

[0099] Optionally, in another implementation, after the category modification in the above step, combining multiple entities of the address category in each entity according to the administrative district level to generate the target address may include the following steps:

[0100] After the category correction is performed, determining whether there is a missing address between any two entities among the multiple entities according to the administrative district levels to which the multiple entities belong;

[0101] When it is determined that there are missing addresses between any two entities, corresponding missing addresses are obtained from the address data set according to the multiple entities and the missing addresses are supplemented between the two entities;

[0102] A target address is generated according to the plurality of entities and the supplemented missing addresses.

[0103] It is understandable that there may be a situation where the address in the query information input by the user is incomplete. Even after the category is corrected, the address may still be incomplete. Therefore, the missing part can be completed based on the corrected address entity.

[0104] For example, the corrected types of the three entities "Gansu", "Baiyin" and "Chang'an Road" are all addresses. These three entities are the administrative levels of province, city and street respectively. Therefore, it can be determined that the district / county level is missing. Therefore, the address entity corresponding to the district / county level can be obtained from the address data set. The supplemented target address can be "Chang'an Road, Baiyin District, Baiyin City, Gansu Province", which is equivalent to adding a new address entity to the information entered by the user, making the address more complete.

[0105] Through the above technical solution, after obtaining the query message, the information in the query message is entity recognized to obtain the categories of each entity in the information. When there is a first entity of address category in the information, the address subset corresponding to the first entity is matched according to the address data set, and when the address subset corresponding to the first entity is matched, the category of entities of other categories located after the first entity in the information is corrected according to the address subset. Therefore, for the query information input by the user, the entity recognition model is first used to identify the entity category, preliminarily determine the entity type in the query information, locate the address entity category, and then the entity after the address entity that is not recognized as an address is matched through the address data set to determine whether to correct the category to an address, which can solve the problem of identifying the address as other entity types, thereby improving the address matching accuracy in address recognition.

[0106] Figure 5 FIG. 1 is a block diagram of an address processing device according to an exemplary embodiment. Figure 5 The address processing device 500 includes an acquisition module 510 , an identification module 520 , a matching module 530 and a processing module 540 .

[0107] The acquisition module 510 is used to acquire the query message;

[0108] The identification module 520 is used to perform entity identification on the information in the query message to obtain the category of each entity of the information;

[0109] The matching module 530 is used to match the address subset corresponding to the first entity according to the address data set when there is a first entity of address type in the information;

[0110] The processing module 540 is used to modify the categories of entities of other categories located after the first entity in the information according to the address subset when the address subset corresponding to the first entity is matched.

[0111] Optionally, the address data set includes address subsets corresponding to multiple addresses belonging to various administrative levels, including provinces, cities, counties or districts, townships or towns, and villages.

[0112] Optionally, the address subset corresponding to each of the multiple addresses includes: multiple addresses and / or address abbreviations belonging to the next administrative district level of the address.

[0113] Optionally, the identification module 520 may be used to:

[0114] The information in the query message is subjected to entity recognition through the entity recognition model to obtain the category of each entity in the information output by the entity recognition model.

[0115] Optionally, the address processing device 500 may further include: an entity determination module, configured to:

[0116] When there are multiple consecutive entities of the address type, the last entity of the multiple consecutive entities is determined as the first entity;

[0117] When there are multiple non-continuous entities of the address type, all of the multiple non-continuous entities are regarded as the first entity.

[0118] Optionally, the entity determination module is further configured to:

[0119] When the last entity belongs to at least two upper-level administrative regions, the last n entities of the plurality of consecutive entities are collectively regarded as the first entity.

[0120] Optionally, the processing module 540 may include:

[0121] A query submodule, for searching, in the case where the address subset corresponding to the first entity is matched, whether there is an address matching the second entity in the address subset corresponding to the first entity, the second entity being an entity located after the first entity and adjacent to the first entity;

[0122] a correction submodule, configured to correct the type of the second entity to an address if there is an address matching the second entity in the address subset corresponding to the first entity;

[0123] The query submodule is further used to query whether there is a corresponding address subset for the second entity in the address data set;

[0124] The execution submodule is used to, when the address subset corresponding to the second entity is queried, take the second entity as a new first entity and repeatedly execute the step of searching the address subset corresponding to the first entity to see whether there is an address matching the second entity.

[0125] Optionally, the processing module 540 may further include: a determination submodule, configured to:

[0126] When there is no address matching the second entity in the address subset corresponding to the first entity, or when the address subset corresponding to the second entity cannot be found, the category modification ends.

[0127] Optionally, the address processing device 500 may further include:

[0128] An address generation module is used to determine a target address corresponding to the information according to the category of each entity after category correction;

[0129] An information acquisition module, used for acquiring at least one piece of relevant enterprise information according to the target address;

[0130] The output module is used to output at least one piece of enterprise information.

[0131] Optionally, the address generation module is used to:

[0132] After the category modification is performed, multiple entities of the address category among the entities are combined according to the administrative district level to generate the target address.

[0133] Optionally, the address generation module is used to:

[0134] After performing the category correction, determining whether there is an address missing between any two entities among the multiple entities according to the administrative district levels to which the multiple entities belong;

[0135] When it is determined that there are missing addresses between any two entities, corresponding missing addresses are obtained from the address data set according to the multiple entities and the missing addresses are supplemented between the two entities;

[0136] A target address is generated according to the plurality of entities and the supplemented missing addresses.

[0137] Through the above technical solution, for the query information input by the user, the entity recognition model is first used to identify the entity category, preliminarily determine the entity type in the query information, locate the address entity category, and then the entities after this address entity that are not recognized as addresses are matched through the address data set to determine whether to correct the category to address. This can solve the problem of identifying addresses as other entity types, thereby improving the address matching accuracy in address recognition.

[0138] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0139] Figure 6 600 is a block diagram of an electronic device according to an exemplary embodiment. For example, the electronic device 600 may be provided as a server. Figure 6 The electronic device 600 includes a processor 622, which may be one or more, and a memory 632 for storing a computer program executable by the processor 622. The computer program stored in the memory 632 may include one or more modules, each corresponding to a set of instructions. In addition, the processor 622 may be configured to execute the computer program to perform the above-mentioned address processing method.

[0140] In addition, the electronic device 600 may further include a power supply component 626 and a communication component 650. The power supply component 626 may be configured to perform power management of the electronic device 600, and the communication component 650 may be configured to implement communication of the electronic device 600, for example, wired or wireless communication. In addition, the electronic device 600 may further include an input / output (I / O) interface 658. The electronic device 600 may operate based on an operating system stored in the memory 632, such as Windows Server 2000. TM , Mac OSX TM , Unix TM , Linux TM etc.

[0141] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the above-mentioned address processing method are implemented. For example, the non-transitory computer-readable storage medium can be the above-mentioned memory 632 including program instructions, and the above-mentioned program instructions can be executed by the processor 622 of the electronic device 600 to complete the above-mentioned address processing method.

[0142] In another exemplary embodiment, a computer program product is further provided. The computer program product includes a computer program executable by a programmable device. The computer program has a code portion for executing the above-mentioned address processing method when executed by the programmable device.

[0143] The preferred embodiments of the present invention are described in detail above in conjunction with the accompanying drawings. However, the present invention is not limited to the specific details in the above embodiments. Within the technical concept of the present invention, a variety of simple modifications can be made to the technical solution of the present invention, and these simple modifications all belong to the protection scope of the present invention.

[0144] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present invention will not further describe various possible combinations.

[0145] In addition, various embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the present invention, they should also be regarded as the contents disclosed by the present invention.

Claims

1. An address processing method, It is characterized in that include: Get query message; Performing entity recognition on the information in the query message to obtain categories of each entity in the information; In the case where a first entity of the address type exists in the information, an address subset corresponding to the first entity is matched according to the address data set; wherein, in the case where there are multiple continuous entities of the address type, the last entity of the multiple continuous entities is determined as the first entity; and in the case where there are multiple non-continuous entities of the address type, all of the multiple non-continuous entities are taken as the first entity; In the case where the address subset corresponding to the first entity is matched, the categories of entities of other categories located after the first entity in the information are modified according to the address subset.

2. The method according to claim 1, It is characterized in that The performing entity recognition on the information in the query message to obtain the category of each entity of the information includes: The information in the query message is subjected to entity recognition through an entity recognition model to obtain the category of each entity in the information output by the entity recognition model.

3. The method according to claim 1, It is characterized in that The determining the last entity of the plurality of consecutive entities as the first entity comprises: In the case that the last entity belongs to at least two upper administrative district levels, the last n entities of the multiple continuous entities are collectively regarded as the first entity.

4. The method according to claim 1, It is characterized in that In the case where the address subset corresponding to the first entity is matched, modifying the category of entities of other categories located after the first entity in the information according to the address subset includes: In the case where the address subset corresponding to the first entity is matched, searching the address subset corresponding to the first entity to see whether there is an address matching a second entity, where the second entity is an entity located after the first entity and adjacent to the first entity; If there is an address matching the second entity in the address subset corresponding to the first entity, correct the type of the second entity to an address; Querying the address data set whether a corresponding address subset exists for the second entity; In the case that the second entity is found to have a corresponding address subset, the second entity is taken as a new first entity, and the step of searching the address subset corresponding to the first entity to see whether there is an address matching the second entity is repeated.

5. The method according to claim 4, It is characterized in that When there is no address matching the second entity in the address subset corresponding to the first entity, or when the address subset corresponding to the second entity cannot be found, the category modification ends.

6. The method according to claim 1, It is characterized in that The method further comprises: Determining a target address corresponding to the information according to the category of each entity after category correction; Acquire at least one related enterprise information according to the target address; The at least one enterprise information is outputted.

7. The method according to claim 6, It is characterized in that The step of determining the target address corresponding to the information according to the category of each entity after category modification includes: After the category modification is performed, multiple entities of the address category among the entities are combined according to the administrative district level to generate the target address.

8. The method according to claim 7, It is characterized in that After the category modification is performed, the entities of the category of address among the entities are combined according to the administrative district level to generate the target address, including: After performing the category correction, determining whether there is an address missing between any two entities among the multiple entities according to the administrative district levels to which the multiple entities belong; When it is determined that there are missing addresses between any two entities, obtaining corresponding missing addresses from the address data set according to the multiple entities and supplementing the missing addresses between the two entities; The target address is generated according to the multiple entities and the supplemented missing address.

9. The method according to any one of claims 1 to 8, It is characterized in that The address data set includes address subsets corresponding to multiple addresses belonging to various administrative levels, and the various administrative levels include province, city, county or district, township or town, and village.

10. The method according to claim 9, It is characterized in that The address subset corresponding to each of the multiple addresses includes: multiple addresses and / or address abbreviations belonging to the next administrative district level of the address.

11. An address processing device, It is characterized in that include: An acquisition module, used to acquire query messages; An identification module, used to perform entity identification on the information in the query message to obtain the category of each entity of the information; a matching module, configured to match, in the case where a first entity of the address category exists in the information, an address subset corresponding to the first entity according to the address data set; wherein, in the case where there are multiple continuous entities of the address category, the last entity of the multiple continuous entities is determined as the first entity; and in the case where there are multiple non-continuous entities of the address category, all of the multiple non-continuous entities are taken as the first entity; The processing module is used to modify the categories of entities of other categories located after the first entity in the information according to the address subset when matching the address subset corresponding to the first entity.

12. A non-transitory computer-readable storage medium having stored thereon a computer program, It is characterized in that When the program is executed by a processor, the steps of the method according to any one of claims 1 to 10 are implemented.

13. An electronic device, It is characterized in that include: a memory having a computer program stored thereon; A processor, configured to execute the computer program in the memory to implement the steps of the method according to any one of claims 1 to 10.