Address data processing method and device, computer device and storage medium

Abstract

The application relates to an address data processing method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring a multi-level standardized address dictionary library; based on the multi-level standardized address dictionary library, a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree are constructed and generated; a target address to be processed is acquired, the target address to be processed is filtered based on the second-granularity multi-level address dictionary tree, and a target recall address is determined; the target recall address is subjected to keyword segmentation based on the first-granularity multi-level address dictionary tree, and an address keyword is obtained; the address keyword is extracted based on the first-granularity multi-level address dictionary tree, a target keyword is obtained, and an intermediate standardized address is generated based on the target keyword; the intermediate standardized address is subjected to address link analysis based on the first-granularity multi-level address dictionary tree, and a target standardized address is determined. The method can improve the search matching efficiency of a character string.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to an address data processing method, apparatus, computer equipment, storage medium, and computer program product. Background Technology

[0002] With the development of word segmentation matching standardization technology, address standardization technology based on address link analysis has emerged. This technology can quickly query and analyze the correlation of matching address links from bottom to top based on the multi-level Chinese address dictionary tree with inverted index and AC automaton search technology. Then, it effectively cleans, corrects and supplements the missing, duplicate, redundant or even erroneous content in the address links, and finally returns standardized six-level address data with high completeness, thereby effectively improving the search and matching efficiency of address strings.

[0003] However, traditional word segmentation matching standardization technology cannot correctly return complete and valid information when faced with problems such as incorrect or missing information in the data. Its dictionary tree search adopts a top-down matching search, which results in low efficiency in searching and matching address strings. Summary of the Invention

[0004] Therefore, it is necessary to provide an address data processing method, apparatus, computer device, computer-readable storage medium, and computer program product that can efficiently search and match address strings, thereby improving the efficiency of address string search and matching.

[0005] An address data processing method, the method comprising:

[0006] Obtain a multi-level standardized address dictionary, which is used to store standardized addresses;

[0007] Based on the multi-level standardized address dictionary library, a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree are constructed and generated, wherein the precision of the second granularity is greater than that of the first granularity.

[0008] Obtain the target address to be processed, filter the target address to be processed based on the second granularity multi-level address dictionary tree, and determine the target recall address;

[0009] Based on the first granular multi-level address dictionary tree, the target recall address is segmented into keywords to obtain address keywords;

[0010] Address keywords are extracted based on the first granularity multi-level address dictionary tree to obtain target keywords, and intermediate standardized addresses are generated based on the target keywords;

[0011] Based on the first granularity multi-level address dictionary tree, address link analysis is performed on the intermediate standardized address to determine the target standardized address.

[0012] In one embodiment, before obtaining a multi-level standardized address dictionary, the method further includes:

[0013] Obtain the aliases and attributes corresponding to each address in the original multi-level standardized address library;

[0014] Each alias and its corresponding attribute are added to the original multi-level standardized address library to obtain the multi-level standardized address dictionary library.

[0015] In one embodiment, based on the multi-level standardized address dictionary library, a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree are constructed and generated, wherein the precision of the second granularity is greater than that of the first granularity, including:

[0016] Obtain the address information corresponding to each level of address in the multi-level standardized address dictionary library, store the address information in the corresponding address block, and generate an address block dictionary based on each address block;

[0017] Create the root node of the first-granularity multi-level address dictionary tree, obtain the address name corresponding to each address information in the address block dictionary, and based on the level corresponding to the address name corresponding to each address information, create nodes corresponding to the level layer by layer starting from the root node. Each node contains the address name corresponding to the level, until the address name corresponding to each address information has a corresponding node, thus generating the first-granularity multi-level address dictionary tree.

[0018] Based on the address block dictionary, obtain the address name and address index number corresponding to each address information, and generate an inverted index dictionary based on the address name and address index number;

[0019] The second granularity multilevel address dictionary tree is generated based on the inverted index dictionary. Each node of the second granularity multilevel address dictionary tree contains a character corresponding to the address name in the inverted index dictionary.

[0020] In one embodiment, before obtaining the target address to be processed, the method further includes:

[0021] Obtain the original address to be processed, remove the first special character from the original address to be processed, and obtain the first address to be processed.

[0022] Calculate the length of each string in the first address to be processed, delete the strings that do not meet the preset length, and obtain the second address to be processed;

[0023] Replace the second special character corresponding to each string in the second address to be processed with a preset character to obtain the third address to be processed;

[0024] The third address to be processed is segmented to obtain the target address to be processed.

[0025] In one embodiment, obtaining the target address to be processed, filtering the target address to be processed based on the second granularity multi-level address dictionary tree, and determining the target recall address includes:

[0026] Starting from the starting character corresponding to the target address to be processed, the target address to be processed is searched and matched based on the second granularity multi-level address dictionary tree to obtain an initial recall address. The initial recall address includes a first recall address and a second recall address, wherein the character length of the first recall address is greater than the character length of the second recall address.

[0027] When the second recall address exists in the first recall address, the second recall address is deleted to obtain the intermediate recall address;

[0028] Delete the addresses obtained by incorrect matching from the intermediate recall addresses to obtain the target recall addresses.

[0029] In one embodiment, address link analysis is performed on the intermediate standardized addresses based on the first granularity multi-level address dictionary tree to determine the target standardized addresses, including:

[0030] Based on the first granularity multi-level address dictionary tree, a bottom-up search and matching is performed on the intermediate standardized addresses, and the number of matching addresses matched by each intermediate standardized address is calculated. The number of matching addresses is then used as the matching length corresponding to the intermediate standardized address.

[0031] Based on the matching length and criteria, a second standardized address is selected from the intermediate standardized addresses, the missing address information of the second standardized address is supplemented, and the corresponding confidence label is assigned to the second standardized address.

[0032] Based on the credibility tag, the target standardized address is obtained from the second standardized address.

[0033] In one embodiment, based on each of the matching lengths and criteria, a second standardized address is selected from the intermediate standardized addresses, missing address information of the second standardized address is supplemented, and a corresponding confidence label is assigned to the second standardized address, including:

[0034] Get the threshold;

[0035] Based on each of the aforementioned matching lengths, the intermediate standardized address with the largest matching length is selected from the intermediate standardized addresses as the third standardized address, and the number of address entries for the third standardized address is calculated, with the number of address entries being used as the first number.

[0036] When the condition is that the first number equals the threshold, the third standardized address is used as the second standardized address to supplement the missing address information of the second standardized address, and the second standardized address is labeled with a first confidence label.

[0037] Calculate the number of address entries with high-level address information in the third standardized address, and use the number of address entries as the second number;

[0038] When the first number is greater than the threshold and the second number is equal to the threshold, the third standardized address with high-level address information is used as the second standardized address to supplement the missing address information of the second standardized address, and the second standardized address is labeled with a first confidence label.

[0039] Obtain addresses with the same level of suffix address information from the third standardized address, select the address with the latest suffix address information as the latest address, calculate the number of addresses of the latest address, and use the number of addresses of the latest address as the third number.

[0040] When the first number and the second number are both greater than the threshold, and the third number is equal to the threshold, the last address is taken as the second standardized address, the missing address information of the second standardized address is supplemented, and the second standardized address is labeled with a second confidence label.

[0041] When the first number, the second number, and the third number are all greater than the threshold, an address is randomly selected from the third standardized address as the second standardized address to supplement the missing address information of the second standardized address, and a third confidence label is marked on the second standardized address. The first confidence label is greater than the second confidence label, and the second confidence label is greater than the third confidence label.

[0042] An address data processing apparatus, the apparatus comprising:

[0043] A multi-level standardized address dictionary acquisition module is used to acquire a multi-level standardized address dictionary, which is used to store standardized addresses;

[0044] The dictionary tree generation module is used to construct and generate a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree based on the multi-level standardized address dictionary library, wherein the precision of the second granularity is greater than that of the first granularity.

[0045] The target recall address determination module is used to obtain the target address to be processed, filter the target address to be processed based on the second granularity multi-level address dictionary tree, and determine the target recall address;

[0046] The address keyword determination module is used to perform keyword segmentation on the target recall address based on the first granular multi-level address dictionary tree to obtain address keywords;

[0047] The intermediate standardized address generation module is used to extract address keywords based on the first granularity multi-level address dictionary tree to obtain target keywords, and generate intermediate standardized addresses based on the target keywords;

[0048] The target standardized address determination module is used to perform address link analysis on the intermediate standardized addresses based on the first granularity multi-level address dictionary tree to determine the target standardized address.

[0049] A computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program performing the following steps:

[0050] Obtain a multi-level standardized address dictionary, which is used to store standardized addresses;

[0051] Based on the multi-level standardized address dictionary library, a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree are constructed and generated, wherein the precision of the second granularity is greater than that of the first granularity.

[0052] Obtain the target address to be processed, filter the target address to be processed based on the second granularity multi-level address dictionary tree, and determine the target recall address;

[0053] Based on the first granular multi-level address dictionary tree, the target recall address is segmented into keywords to obtain address keywords;

[0054] Address keywords are extracted based on the first granularity multi-level address dictionary tree to obtain target keywords, and intermediate standardized addresses are generated based on the target keywords;

[0055] Based on the first granularity multi-level address dictionary tree, address link analysis is performed on the intermediate standardized address to determine the target standardized address.

[0056] A computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, performs the following steps:

[0057] Obtain a multi-level standardized address dictionary, which is used to store standardized addresses;

[0058] Based on the multi-level standardized address dictionary library, a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree are constructed and generated, wherein the precision of the second granularity is greater than that of the first granularity.

[0059] Obtain the target address to be processed, filter the target address to be processed based on the second granularity multi-level address dictionary tree, and determine the target recall address;

[0060] Based on the first granular multi-level address dictionary tree, the target recall address is segmented into keywords to obtain address keywords;

[0061] Address keywords are extracted based on the first granularity multi-level address dictionary tree to obtain target keywords, and intermediate standardized addresses are generated based on the target keywords;

[0062] Based on the first granularity multi-level address dictionary tree, address link analysis is performed on the intermediate standardized address to determine the target standardized address.

[0063] A computer program product. The computer program product includes a computer program that, when executed by a processor, performs the following steps:

[0064] Obtain a multi-level standardized address dictionary, which is used to store standardized addresses;

[0065] Based on the multi-level standardized address dictionary library, a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree are constructed and generated, wherein the precision of the second granularity is greater than that of the first granularity.

[0066] Obtain the target address to be processed, filter the target address to be processed based on the second granularity multi-level address dictionary tree, and determine the target recall address;

[0067] Based on the first granular multi-level address dictionary tree, the target recall address is segmented into keywords to obtain address keywords;

[0068] Address keywords are extracted based on the first granularity multi-level address dictionary tree to obtain target keywords, and intermediate standardized addresses are generated based on the target keywords;

[0069] Based on the first granularity multi-level address dictionary tree, address link analysis is performed on the intermediate standardized address to determine the target standardized address.

[0070] The aforementioned address data processing method, apparatus, computer equipment, storage medium, and computer program product involve: acquiring a multi-level standardized address dictionary library, which stores standardized addresses; constructing a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree based on the multi-level standardized address dictionary library, wherein the precision of the second granularity is greater than that of the first granularity; acquiring a target address to be processed; filtering the target address to be processed based on the second-granularity multi-level address dictionary tree to determine a target recall address; performing keyword segmentation on the target recall address based on the first-granularity multi-level address dictionary tree to obtain address keywords; extracting target keywords based on the address keywords based on the first-granularity multi-level address dictionary tree; generating intermediate standardized addresses based on the target keywords; and performing address link analysis on the intermediate standardized addresses based on the first-granularity multi-level address dictionary tree to determine the target standardized address. By acquiring a multi-level standardized address dictionary, a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree are constructed. Based on the second-granularity multi-level address dictionary tree, target addresses to be processed are filtered to determine target recall addresses. Based on the first-granularity multi-level address dictionary tree, keyword segmentation and extraction are performed on the target recall addresses to obtain target keywords. Intermediate standardized addresses are generated based on the target keywords. Address link analysis is performed on the intermediate standardized addresses based on the first-granularity multi-level address dictionary tree to determine target standardized addresses. The first-granularity multi-level address dictionary tree and the second-granularity multi-level address dictionary tree search and match address strings from bottom to top, thereby improving the search and matching efficiency of address strings. Attached Figure Description

[0071] Figure 1 This is an application environment diagram of an address data processing method in one embodiment;

[0072] Figure 2 This is a flowchart illustrating an address data processing method in one embodiment;

[0073] Figure 3 A flowchart illustrating the process of reading from a multi-level standardized address dictionary.

[0074] Figure 4 Here is an example graph of a bottom-up search of the dictionary tree in an instance;

[0075] Figure 5 This is a flowchart illustrating the address link analysis process.

[0076] Figure 6 A flowchart illustrating the standardization and rapid resolution process for Chinese addresses;

[0077] Figure 7 A flowchart illustrating the process of adding address aliases in one embodiment;

[0078] Figure 8 This is a schematic diagram of the dictionary tree generation process in one embodiment;

[0079] Figure 9 This is a schematic diagram of the structure of a first-level multi-granularity address dictionary tree;

[0080] Figure 10 A flowchart illustrating the process of processing the original address to be processed;

[0081] Figure 11 This is a schematic diagram of the process for determining the target recall address in one embodiment;

[0082] Figure 12 This is a schematic diagram of the process for determining the target standardized address in one embodiment;

[0083] Figure 13 This is a flowchart illustrating the process of labeling a second standardized address with a credibility tag in one embodiment.

[0084] Figure 14 This is a structural block diagram of an address data processing device in one embodiment;

[0085] Figure 15 This is an internal structural diagram of a computer device in one embodiment;

[0086] Figure 16 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0087] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0088] The address data processing method provided in this application embodiment can be applied to, for example, Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located on the cloud or other network servers. Terminal 102 is used to input and display address data, which includes standardized and non-standardized address data. Server 104 is used to determine the target standardized address. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. Server 104 can be implemented using a standalone server or a server cluster consisting of multiple servers.

[0089] In one embodiment, such as Figure 2 As shown, an address data processing method is provided, which is applied to... Figure 1 Taking the server in the example, the following steps are included:

[0090] Step S200: Obtain a multi-level standardized address dictionary, which is used to store standardized addresses.

[0091] Among them, the multi-level standardized address dictionary refers to an information database used to store address-related information. It is not limited to storing standardized addresses, but also includes address aliases, address abbreviations, special names, and other information.

[0092] Specifically, the database initially stores collected standardized address data. However, since addresses also have aliases, abbreviations, special designations, and other information, and these can be used to find corresponding standardized addresses, the collected information on historical address versions, abbreviations, and special designations is added to the original multi-level standardized address dictionary. By saving address version snapshots, historical versions, abbreviations, and special designations are recorded, ensuring that the desired address can be found using historical versions, abbreviations, and special designations. The original multi-level standardized address dictionary is updated periodically by modifying the address-related data to obtain a multi-level standardized address dictionary, preparing data for standardizing non-standardized addresses. The original multi-level standardized address dictionary refers to the address information database before newly collected address information is added, and this address information is not limited to historical address versions, abbreviations, and special designations.

[0093] Step S202: Based on the multi-level standardized address dictionary library, construct and generate a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree, wherein the precision of the second granularity is greater than that of the first granularity.

[0094] Wherein, the first-granularity multi-level address dictionary tree refers to a trie for storing six-level administrative division information, each node of which stores information of one level of administrative division. For example, after a node stores "Guangdong Province", its child nodes can be a node storing "Guangzhou City", a node storing "Shenzhen City", etc.; and the child nodes of the node storing "Shenzhen City" can be a node storing "Nanshan District", a node storing "Futian District", etc., and so on for other nodes. The second-granularity multi-level address dictionary tree refers to a trie of AC automaton in which each node stores a single character, and it is a trie in character string order. For example, after a node stores the Chinese character "广", its child nodes can be a node storing the Chinese character "东", a node storing the Chinese character "西", a node storing the Chinese character "州", etc. The AC automaton refers to a combination of the KMP algorithm and a Trie tree, which is suitable for solving multi-pattern string matching, that is, can solve the problem of finding multiple candidate substrings in a long character string. In one example of the present application, the AC automaton is used to match information of levels including province, city, district, street, road and building in one search, achieving the effect of high search efficiency. The KMP algorithm is a single-pattern string matching algorithm, which handles the problem of searching for a candidate substring in a long character string. Single-pattern string matching refers to a matching mode for checking whether one character string contains another character string. Multi-pattern string matching refers to a matching mode for matching multiple character strings (key character strings) from a segment of character string, that is, finding all substrings identical to a certain substring in a character string. A Trie tree refers to a dictionary tree, which is a tree structure that can quickly search whether a group of character strings contains a character string.

[0095] Specifically, based on existing address data in a multi-level standardized dictionary library, a second-granularity multi-level address dictionary tree for obtaining address-related character strings from unstandardized disordered address data, and a first-granularity multi-level address dictionary tree for analyzing association information between the address-related character strings to convert the character strings into standardized addresses based on the association information are constructed. Wherein, the standardized addresses in the multi-level standardized dictionary library are standardized according to the level order of province, city, district, street, building, etc. Before constructing the first-granularity multi-level address dictionary tree and the second-granularity multi-level address dictionary tree, it is necessary to store address information in the multi-level standardized dictionary library into address blocks, save the address blocks in a dictionary, and then construct the first-granularity multi-level address dictionary tree and the second-granularity multi-level address dictionary tree based on the dictionary information. Wherein, the process of reading the multi-level standardized address dictionary library is as Figure 3 shown, wherein in the figure, id refers to a node number corresponding to a node during programming, last_id refers to a current node number during programming, RegionEntity refers to an address block for storing address information, and ad_map refers to a data structure of each address block in the administrative division trie, which is a dictionary, that is, the data structure of each address block in the first-granularity multi-level address dictionary tree.

[0096] Step S204: Obtain the target address to be processed, filter the target address to be processed based on the second granularity multi-level address dictionary tree, and determine the target recall address.

[0097] Here, the target address to be processed refers to the address characters obtained by cleaning and segmenting the address data before any processing. The target recall address refers to the address string obtained after searching and matching the target address to be processed through a second-granularity multi-level address dictionary tree and then filtering it.

[0098] Specifically, the collected raw address data contains many special characters, which have a significant negative impact on the standardization of the raw address data. Therefore, before searching and matching the raw address data through the second-granularity multi-level address dictionary tree and the first-granularity multi-level address dictionary tree, it is necessary to perform address validity checks and remove special characters from the address data to reduce the impact of invalid character information on address standardization. Specifically, during address validity checks, strings that do not contain valid address information are deleted. When removing special strings from the address data, special characters, punctuation characters, problematic characters, etc., are replaced or deleted. Furthermore, full-width strings need to be converted to half-width strings, and Chinese numerals need to be converted to Arabic numerals.

[0099] Step S206: Based on the first granular multi-level address dictionary tree, the target recall address is segmented into keywords to obtain address keywords.

[0100] Here, "keywords" refers to words in the target recall address that are identical or similar to the address names in the first-level multi-level address dictionary. "Address keywords" refers to all words obtained after keyword segmentation of the target recall address, including words that are identical or similar to the address names and words that are unrelated to the address names.

[0101] Specifically, the target recall address is a long string, and each string may contain multiple non-standard address strings. It is necessary to perform keyword segmentation on the target recall address to separate the various address names contained in the string, thereby improving the efficiency of subsequent search and matching by the first-granularity multi-level address dictionary tree.

[0102] Step S208: Extract address keywords based on the first granular multi-level address dictionary tree to obtain target keywords, and generate intermediate standardized addresses based on the target keywords.

[0103] Here, target keywords refer to words selected from address keywords that are identical or similar to address names in the first-level multi-level address dictionary. Intermediate standardized addresses refer to addresses that are standardized by combining target keywords using the first-level multi-level address dictionary and the relationships between target keywords, and then selecting addresses that conform to the order of province, city, district, street, road, building, etc. Association information refers to the distance between target keywords. For example, if Guangzhou is closer to the phrase "Guangdong Province" than to the phrase "Guangxi Zhuang Autonomous Region," then Guangzhou should be more likely to be combined with Guangdong Province as a keyword.

[0104] Specifically, address keyword extraction requires referring to the address names corresponding to each node in the first-level multi-level address dictionary tree. After obtaining the target keywords, they need to be combined to form standardized addresses, which are addresses combined according to levels such as province, city, district, street, road, and building. When address information for a relevant level is missing, that level can be set to empty, and then keyword combinations are performed for other levels of addresses. Finally, addresses that conform to the standardized addresses are selected from the combined addresses.

[0105] Step S210: Based on the first granularity multi-level address dictionary tree, perform address link analysis on the intermediate standardized address to determine the target standardized address.

[0106] Address link analysis refers to the process of analyzing the number of actual matching address names in intermediate standardized addresses, completing missing information and correcting errors in intermediate standardized addresses, and labeling the completed addresses with credibility tags. Target standardized addresses refer to addresses that meet address standardization requirements and have been completed or corrected.

[0107] Specifically, based on the first-granularity multi-level address dictionary tree, each address name is searched backwards from the lower-level address (e.g., building) to the higher-level address (e.g., province) of the intermediate standardized address. When a corresponding address name is missing or incorrect, information completion and error correction are performed. The number of address names actually matched in the intermediate standardized address is analyzed, and this number is taken as the actual matching length of the intermediate standardized address. The intermediate standardized address corresponding to the longest matching length is selected as the result with the highest confidence. Then, the intermediate standardized address is returned as the target standardized address. After obtaining the target standardized address, the corresponding latitude and longitude can be mapped according to the target standardization, and the latitude and longitude data can be used for the required business scenarios. One example of bottom-up search of the dictionary tree is... Figure 4 As shown, Figure 4The path highlighted in bold with a reverse direction arrow is one example of a bottom-up search during address link analysis. The address link analysis process is as follows: Figure 5 As shown, Figure 5 In this diagram, S, A, and B represent different levels of trustworthiness labels, where S > A > B. The process of standardizing and quickly resolving Chinese addresses, as shown in the flowchart, illustrates one example of the process from non-standardized to standardized addresses. Figure 6 .

[0108] The aforementioned address data processing method, apparatus, computer equipment, storage medium, and computer program product involve: acquiring a multi-level standardized address dictionary library, which stores standardized addresses; constructing a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree based on the multi-level standardized address dictionary library, wherein the precision of the second granularity is greater than that of the first granularity; acquiring a target address to be processed; filtering the target address to be processed based on the second-granularity multi-level address dictionary tree to determine a target recall address; performing keyword segmentation on the target recall address based on the first-granularity multi-level address dictionary tree to obtain address keywords; extracting target keywords based on the address keywords based on the first-granularity multi-level address dictionary tree; generating intermediate standardized addresses based on the target keywords; and performing address link analysis on the intermediate standardized addresses based on the first-granularity multi-level address dictionary tree to determine the target standardized address. By acquiring a multi-level standardized address dictionary, a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree are constructed. Based on the second-granularity multi-level address dictionary tree, target addresses to be processed are filtered to determine target recall addresses. Based on the first-granularity multi-level address dictionary tree, keyword segmentation and extraction are performed on the target recall addresses to obtain target keywords. Intermediate standardized addresses are generated based on the target keywords. Address link analysis is performed on the intermediate standardized addresses based on the first-granularity multi-level address dictionary tree to determine target standardized addresses. The first-granularity multi-level address dictionary tree and the second-granularity multi-level address dictionary tree search and match address strings from bottom to top, thereby improving the search and matching efficiency of address strings.

[0109] In one embodiment, such as Figure 7 As shown, before step S200, the following steps are also included:

[0110] Step S700: Obtain the aliases corresponding to each address in the original multi-level standardized address library and the attributes corresponding to the aliases.

[0111] The original multi-level standardized address database refers to an address information database that has not yet added newly collected address information. This address information is not limited to historical address versions, abbreviations, or special names. Aliases refer to information including, but not limited to, historical address versions, abbreviations, and special names related to the address. Aliase attributes refer to attribute information related to the alias.

[0112] Specifically, in order to ensure that the corresponding address information can be found even when using aliases, it is necessary to periodically add address-related address information to the original multi-level standardized address database and update the address information.

[0113] Step S702: Add each of the aliases and their corresponding attributes to the original multi-level standardized address library to obtain the multi-level standardized address dictionary library.

[0114] Specifically, adding aliases and alias attributes to the original multi-level normalized address library not only ensures that the corresponding address information can be found using aliases, but also enriches the multi-level normalized address dictionary, thereby providing more effective information about addresses and improving the efficiency of the dictionary tree's search and matching.

[0115] In this embodiment, by periodically collecting and adding address aliases and their corresponding attributes to the original multi-level standardized address database, a multi-level standardized address dictionary database is obtained. This ensures that the corresponding address information can be found using the alias and provides more effective information about the address, thereby improving the efficiency of dictionary tree search matching.

[0116] In one embodiment, such as Figure 8 As shown, step S202 includes:

[0117] Step S800: Obtain the address information corresponding to each level of address in the multi-level standardized address dictionary library, store the address information in the corresponding address block, and generate an address block dictionary based on each address block.

[0118] Address information includes, but is not limited to, address number, parent address number, address type, Chinese name of the address, next-level address number, and address alias. An address block refers to a storage block that stores address information. An address block stores only the attribute information of one administrative division. For example, "Guangdong Province" is an address block, "Guangzhou City" is an address block, and "Tianhe District" is an address block. An address block may retain multiple attribute information. For example, the parent address block of the "Guangzhou City" address block is "Guangdong Province," and its child address blocks are "Tianhe District," "Baiyun District," "Yuexiu District," etc., and so on for other address blocks. An address block dictionary refers to a dictionary in programming, which is an unordered, mutable, and indexed collection. In Python, dictionaries are written in curly braces and have keys and values. In one instance of this application, the dictionary key is the corresponding number of the address block, and the value is the address block itself. For example, the key of the address block "Guangdong Province" is 440, and the value is the index pointer address (the computer's storage address) where this address block is stored.

[0119] Specifically, before constructing the dictionary tree, the address information in the multi-level standardized address dictionary library needs to be implemented into the data structure of the dictionary tree. Then, based on the data structure, the corresponding address block dictionary is constructed. The address block dictionary provides data basis for the construction of the first-granularity multi-level address dictionary tree and the second-granularity multi-level address dictionary tree in the subsequent process.

[0120] Step S802: Create the root node of the first granularity multi-level address dictionary tree, obtain the address name corresponding to each address information in the address block dictionary, and based on the level corresponding to the address name corresponding to each address information, create nodes corresponding to the level layer by layer starting from the root node. Each node contains the address name corresponding to the level, until a corresponding node has been created for each address name corresponding to each address information, thereby generating the first granularity multi-level address dictionary tree.

[0121] Here, the root node refers to the root node corresponding to the first-granularity multi-level address dictionary tree, and the root node does not contain any arbitrary characters. The level refers to the level divided according to the order of province, city, district, street, road, building, etc. In this example, the province is the highest level and the building is the lowest level. That is, the child node of the province corresponds to the city, the child node of the city corresponds to the district, and so on, to construct the first-granularity multi-level address dictionary tree of the corresponding level.

[0122] Specifically, the first-granularity multi-level address dictionary tree is a dictionary tree corresponding to multi-level administrative divisions. Each node stores the address name corresponding to an address block. Each level of the first-granularity multi-level address dictionary tree corresponds to the address name of a first-level administrative division. In addition, to avoid the possibility of duplicate names for different city districts, the address index number corresponding to the address name needs to be added to the corresponding address name. In one example, the structure of the first-granularity multi-level address dictionary tree is as follows: Figure 9 As shown.

[0123] Step S804: Based on the address block dictionary, obtain the address name and address index number corresponding to each address information, and generate an inverted index dictionary based on the address name and address index number.

[0124] Here, the address index number refers to the number corresponding to the address block where the address name is located. The inverted index dictionary refers to a dictionary used to store, but is not limited to, address names, address index numbers, and the second-granularity multi-level address dictionary tree generated in subsequent processes.

[0125] Specifically, constructing a second-granularity multi-level address dictionary tree requires corresponding dictionary data. The corresponding address information is found by searching the address blocks stored in the dictionary, and the second-granularity multi-level dictionary tree is constructed based on the found address information.

[0126] Step S806, generating the second-granularity multi-level address dictionary tree based on the inverted index dictionary, wherein each node corresponding to the second-granularity multi-level address dictionary tree comprises a character corresponding to an address name in the inverted index dictionary.

[0127] Specifically, after finding the information required for constructing the second-granularity multi-level address dictionary tree through the inverted index dictionary, it is also necessary to store the information in the form of individual characters into corresponding nodes. The path passing from the root node to the current node forms an address string, which corresponds to an address name of a certain level. For example, the child nodes of the character "Guang" can be characters such as "Dong", "Xi" and "Zhou" that are geographically associated with the character "Guang", that is, the second-granularity multi-level address dictionary tree is a string-ordered trie.

[0128] In this embodiment, an address block dictionary and an inverted index dictionary are generated through a multi-level standardized dictionary library, and a first-granularity multi-level standardized address dictionary tree and a second-granularity multi-level standardized address dictionary tree are constructed and generated based on the address block dictionary and the inverted index dictionary. The data structure form of trie is applied to store and extract address information, which improves the efficiency of address information search and effectively promotes the search and matching of the trie.

[0129] In one embodiment, as Figure 10 shows, before step S204, the method further includes:

[0130] Step S1000, acquiring an original address to be processed, and removing first special characters from the original address to be processed to obtain a first address to be processed.

[0131] Wherein, the original address to be processed refers to data that has not undergone any processing on the collected original address data. The first special characters refer to non-bracket special characters, including but not limited to "\r\n\t,,。’.`;:、!@??￥$%*^~=+&\”|_\\ / " and the like. The first address to be processed refers to address data remaining after deleting the first special characters in the original address to be processed.

[0132] Specifically, since the original address to be processed has not undergone any processing, it is too messy. Therefore, before performing the subsequent address standardization process, special characters are cleaned once to reduce the analysis and calculation pressure of the subsequent process. However, in order to maintain the integrity of effective address information in the subsequent standardization process, no string content is cleaned or changed in this process. Instead, the original string is retained and the first special characters are filtered out.

[0133] Step S1002, calculating the length of each character string in the first address to be processed, deleting character strings that do not meet a preset length, and obtaining a second address to be processed.

[0134] Here, the preset length refers to a manually defined string length. The second address to be processed refers to the address data remaining after removing strings that do not meet the preset length from the first-generation processed addresses.

[0135] Specifically, the length of the characters is monitored by using the len() function to calculate the string length in programming. Strings with a length less than the preset length in the first address to be processed are deleted to obtain the second address to be processed.

[0136] Step S1004: Replace the second special character corresponding to each string in the second address to be processed with a preset character to obtain the third address to be processed.

[0137] The second special character includes, but is not limited to, special symbols such as "{}【】()()<>\[\]" and "''". Preset symbols refer to various symbols that are manually modified from the second special character, including but not limited to Arabic numerals and half-width strings.

[0138] Specifically, the process begins by formally cleaning and changing the special characters in the second address to be processed. This involves deleting the second special character from the second address to be processed and replacing the deleted special character with an empty string, converting the full-width string in the second address to be processed to a half-width string, and converting the Chinese numerals in the second address to be processed to Arabic numerals, etc., to obtain an address string without the second special character. This address string is then used as the third address to be processed.

[0139] Step S1006: Perform word segmentation on the third address to be processed to obtain the target address to be processed.

[0140] The target address to be processed refers to the address data obtained by performing the deletion of the first special character, the deletion and replacement of the second special character, and word segmentation on the original address to be processed.

[0141] Specifically, the obtained third address to be processed is a long string of addresses. In order to better search and match the address data based on the second-granularity multi-level address dictionary tree, the third address to be processed also needs to be segmented.

[0142] In this embodiment, by performing deletion of the first special character, deletion and replacement of the second special character, and word segmentation on the original address to be processed, the purity of the effective address information of the target address to be processed is higher, and the correlation between the address information is clearer, which greatly improves the efficiency of subsequent search and matching of the first-granularity multi-level address dictionary tree and the second-granularity multi-level address dictionary tree.

[0143] In one embodiment, such as Figure 11 As shown, step S204 includes:

[0144] Step S1100: Starting from the starting character corresponding to the target address to be processed, search and match the target address to be processed based on the second granularity multi-level address dictionary tree to obtain an initial recall address. The initial recall address includes a first recall address and a second recall address. The character length of the first recall address is greater than the character length of the second recall address.

[0145] Here, "search matching" refers to the operation of searching the dictionary tree in a bottom-up order for strings that match the substring of the searched string. "Initial recall address" refers to the address string that matches the original address name, obtained by searching and matching the target address through the first-granularity multi-level address dictionary tree. "First recall address" refers to an initial recall address where the string length is longer and no address is contained within another address. "Second recall address" refers to an initial recall address where the string length is shorter and no address is contained within another address.

[0146] Specifically, starting from the first character of the target address to be processed, each matching address string is searched using the maximum length priority method to avoid sub-address strings contained in the address string. The second recall address may contain sub-address strings from the first recall address.

[0147] Step S1102: When the second recall address exists in the first recall address, the second recall address is deleted to obtain the intermediate recall address.

[0148] The intermediate recall address refers to the address string remaining after deleting the second recall address in the initial address that matches the sub-address string in the first recall address.

[0149] Specifically, the second recall address, which has a sub-address string that matches the first recall address, is deleted to obtain the intermediate recall address. This operation is performed primarily to prevent duplicate words from being segmented between the second and first recall addresses during subsequent word segmentation, thus avoiding unnecessary operations and reducing server load.

[0150] Step S1104: Delete the addresses obtained by matching in the wrong order from the intermediate recall addresses to obtain the target recall address.

[0151] The target recall address refers to the address string obtained after searching and matching the target address to be processed based on the second-granularity multi-level address dictionary tree to obtain the corresponding address string, and then deleting the address strings that do not match.

[0152] Specifically, under normal circumstances, address strings are searched and matched in order of province, city, district, street, road, building, etc. In order to avoid the impact of strings obtained by incorrect search and matching on subsequent processes, it is necessary to remove address strings obtained by violating the order of search and matching in the intermediate recall addresses in order to obtain the target recall address.

[0153] In this embodiment, the target address to be processed is searched and matched based on the second-granularity multi-level address dictionary tree, and the address strings that do not match the search and match are deleted to obtain the target recall address. The second-granularity multi-level address dictionary tree performs matching search in a bottom-up search method, which improves the search efficiency of the dictionary tree.

[0154] In one embodiment, such as Figure 12 As shown, step S210 includes:

[0155] Step S1200: Based on the first granularity multi-level address dictionary tree, perform bottom-up search matching on the intermediate standardized addresses, calculate the number of matching addresses matched by each intermediate standardized address, and use the number of matching as the matching length corresponding to the intermediate standardized address.

[0156] Here, "bottom-up" refers to the order of ascending from the bottom layer of the dictionary tree. "Match count" refers to the number of address names in the intermediate standardized address set that match the address names of each level of administrative division within the first-granularity multi-level address dictionary tree. "Match length" refers to the number of address names in the intermediate standardized address set that match the address names of each level of administrative division within the first-granularity multi-level address dictionary tree.

[0157] Specifically, from low-level addresses to high-level addresses of intermediate standardized addresses, a search and matching process is performed based on the first-granularity multi-level address dictionary tree. The number of matched address names is recorded, and this number is used as the matching length of the corresponding intermediate standardized address to prepare data for subsequent labeling of credibility tags.

[0158] Step S1202: Based on the matching length and criteria, select the second standardized address from the intermediate standardized addresses, supplement the missing address information of the second standardized address, and assign the corresponding confidence label to the second standardized address.

[0159] The criteria used to determine the credibility label type for the second standardized address are as follows: The second standardized address is the address selected from the intermediate standardized addresses when the criteria are met. The credibility label is a tag used to indicate the credibility of the address after search matching, supplementation, or correction.

[0160] Specifically, there may be multiple intermediate standardized addresses. Selecting the target standardized label with the highest credibility requires further analysis. Since the matching lengths and missing or erroneous address information vary among the intermediate standardized addresses, we analyze cases with different matching lengths, different missing address information locations but the same matching length, and the same missing address information location and matching length. Based on the analysis, we assign credibility labels to the selected intermediate standardized addresses to prepare data for the subsequent selection of target standardized addresses.

[0161] Step S1204: Based on the credibility tag, obtain the target standardized address from the second standardized address.

[0162] Specifically, the address with the highest credibility label is selected as the target standardized address, and then the corresponding latitude and longitude can be mapped according to the target standardization result to provide data for application scenarios that require latitude and longitude.

[0163] In this embodiment, intermediate standardized addresses are searched and matched using a first-granularity multi-level address dictionary tree to obtain the matching length corresponding to each intermediate standardized address and the location of missing address information. Based on the matching length and the missing address information, the missing information of the intermediate standardized addresses is supplemented and confidence labels are added. Finally, the target standardized address is selected based on the confidence labels. This method can improve the standardization of addresses and enrich the address information, thus ensuring the accuracy of the obtained target standardized address.

[0164] In one embodiment, such as Figure 13 As shown, step S1202 includes:

[0165] Step S1300: Obtain the threshold.

[0166] The threshold refers to a value set manually; in this application, the threshold is equal to 1.

[0167] Step S1302: Based on each of the matching lengths, select the intermediate standardized address with the largest matching length from the intermediate standardized addresses as the third standardized address, calculate the number of address entries of the third standardized address, and use the number of address entries as the first number.

[0168] The third normalized address refers to the address with the longest matching length among the intermediate normalized addresses. The first number refers to the number of addresses in the third normalized address with the longest matching length.

[0169] Specifically, there may be more than one third standardized address with the longest matching length. Before supplementing the missing address information and labeling the credibility, it is necessary to calculate the number of the third standardized addresses to prepare data for subsequent processes.

[0170] Step S1304: When the condition is that the first number is equal to the threshold, the third standardized address is used as the second standardized address to supplement the missing address information of the second standardized address, and the second standardized address is labeled with a first confidence label.

[0171] The first credibility label refers to the label type corresponding to the highest credibility. In this application, S represents the first credibility label.

[0172] Specifically, when there is only one third standardized address with the longest matching length, the third standardized address is the second standardized address. According to the order of the first granularity multi-level address dictionary tree from the lowest level address to the highest level address, the missing address information in the second standardized address is analyzed and supplemented. When the missing address information is the lowest level address information, the corresponding lowest level address information cannot be found. The lowest level address information needs to be set to empty, and the second standardized address after supplementing the missing information is marked with the first confidence label.

[0173] Step S1306: Calculate the number of address entries with high-level address information in the third standardized address, and use the number of address entries as the second number.

[0174] Here, "high-level address" refers to the address corresponding to the highest level among all the third-standardized addresses, where address information exists at each of the last three levels. "Second number" refers to the number of address entries containing address information among the high-level addresses of the third-standardized addresses when multiple third-standardized addresses have the same maximum matching length.

[0175] Specifically, when the number of addresses corresponding to the third normalized address with the maximum matching length is greater than 1, it is necessary to further judge the addresses with address information in the higher-level addresses of the third normalized address, calculate the number of addresses with address information in the higher-level addresses of the third normalized address, and prepare data for the judgment of subsequent processes.

[0176] Step S1308: When the first number is greater than the threshold and the second number is equal to the threshold, the third standardized address with high-level address information is used as the second standardized address to supplement the missing address information of the second standardized address, and the second standardized address is labeled with a first confidence label.

[0177] Specifically, when there is only one address with address information in the high-level addresses of the third standardized address, the address is taken as the second standardized address. According to the bottom-up order of the first granularity multi-level address dictionary tree, the missing address information in the second standardized address is analyzed and supplemented. When the missing address information is the lowest level address information, the address information corresponding to the lowest level address cannot be found. The lowest level address information needs to be set to empty, and the second standardized address after supplementing the missing information is marked with the first confidence label.

[0178] Step S1310: Obtain addresses with the same level of suffix address information from the third standardized addresses, select the address with the latest suffix address information as the latest address, calculate the number of addresses of the latest address, and use the number of addresses of the latest address as the third number.

[0179] Here, "last address" refers to the address corresponding to the last address string in each third-normalized address. "Last address" refers to the actual position of the third-normalized address with last address information further back in the list. "Third number" refers to the number of address entries in the third-normalized address list that have the same level of last address information as the last address.

[0180] Specifically, when the number of addresses at the last address obtained by calculation is greater than 1, it is necessary to further judge the last address in the third normalized address, calculate the number of addresses at the last address in the third normalized address, and prepare data for the judgment of subsequent processes.

[0181] Step S1312: When the first number and the second number are both greater than the threshold and the third number is equal to the threshold, the last address is taken as the second standardized address, the missing address information of the second standardized address is supplemented, and the second standardized address is labeled with a second confidence label.

[0182] The second credibility label refers to the label type corresponding to a medium level of credibility. In this application, A represents the second credibility label.

[0183] Specifically, when there is only one last address, that is, when the third number is equal to 1, the address is taken as the second standardized address. Based on the search order from the bottom layer to the top layer of the first granularity multi-level address dictionary tree, the missing address information in the second standardized address is analyzed and the missing address information is supplemented. When the missing address information is the lowest level address information, the address information corresponding to the lowest level address cannot be found. The lowest level address information needs to be set to empty, and the second standardized address after supplementing the missing information is marked with a second confidence label.

[0184] Step S1314: When the first number, the second number, and the third number are all greater than the threshold, randomly select an address from the third standardized address as the second standardized address, supplement the missing address information of the second standardized address, and label the second standardized address with a third confidence label, wherein the first confidence label is greater than the second confidence label, and the second confidence label is greater than the third confidence label.

[0185] The third credibility label refers to the label type corresponding to the lowest credibility. In this application, B represents the third credibility label.

[0186] Specifically, when the first number, the second number, and the third number are all greater than the threshold, an address needs to be randomly selected from the intermediate standardized addresses. The selected address is used as the second standardized address. Based on the bottom-up search order of the first granularity multi-level address dictionary tree, the missing address information in the second standardized address is supplemented. When the missing address information is the lowest level address information, the address information corresponding to the lowest level address cannot be found. The lowest level address information needs to be set to empty, and the second standardized address after supplementing the missing information is labeled with a third confidence label.

[0187] In this embodiment, the number of address entries under different conditions is calculated by the matching length of each intermediate standardized address. Based on the criteria corresponding to the number of address entries, the corresponding second standardized address is selected. The missing address information in the second standardized address is supplemented or corrected based on the first granularity multi-level address dictionary tree. The second standardized address after supplementing the missing information is labeled with the corresponding confidence label, which provides a selection basis for the selection of the target standardized address. The bottom-up search method of the first granularity multi-level address dictionary tree improves the search and matching efficiency of address strings.

[0188] In one embodiment, user address data collected from one domain is used as the original address to be processed. The original address to be processed undergoes deletion of a first special character, deletion and replacement of a second special character, and word segmentation to obtain the target address to be processed. A first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree, constructed from the multi-level standardized address dictionary library of the domain, are obtained. The target address to be processed is filtered based on the second-level address dictionary tree to determine the target recall address. The target recall address is segmented into keywords based on the first-granularity multi-level address dictionary tree to obtain address keywords. The address keywords are extracted based on the first-granularity multi-level address dictionary tree to obtain target keywords. The target keywords are then recombine to generate standard multi-level addresses. Addresses conforming to the hierarchical order of province, city, district, street, road, and building are selected from the standard multi-level addresses and used as intermediate standardized addresses. Address link analysis is performed on the intermediate standardized addresses based on the first-granularity multi-level address dictionary tree to determine the target standardized address. According to this method, messy non-standardized addresses can be quickly parsed and standardized address information can be generated from them. Furthermore, when address data contains errors or omissions, the described method can be used to analyze address links, supplementing and correcting missing address information to return more complete address information. It can also map the latitude and longitude information of the corresponding address based on the generated standardized address information, making this latitude and longitude information usable for different business scenarios. This effectively leverages the bottom-up search function of the dictionary tree, significantly improving the efficiency of address string search and matching.

[0189] Based on the same inventive concept, this application also provides an address data processing apparatus for implementing the address data processing method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more address data processing apparatus embodiments provided below can be found in the limitations of the address data processing method described above, and will not be repeated here.

[0190] In one embodiment, such as Figure 14 As shown, an address data processing device is provided, including: a multi-level standardized address dictionary database acquisition module 1400, a dictionary tree generation module 1402, a target recall address determination module 1404, an address keyword determination module 1406, an intermediate standardized address generation module 1408, and a target standardized address determination module 1410, wherein:

[0191] The multi-level standardized address dictionary acquisition module 1400 is used to acquire a multi-level standardized address dictionary, which is used to store standardized addresses.

[0192] The dictionary tree generation module 1402 is used to construct and generate a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree based on the multi-level standardized address dictionary library, wherein the precision of the second granularity is greater than that of the first granularity.

[0193] The target recall address determination module 1404 is used to obtain the target address to be processed, filter the target address to be processed based on the second granularity multi-level address dictionary tree, and determine the target recall address.

[0194] The address keyword determination module 1406 is used to perform keyword segmentation on the target recall address based on the first granularity multi-level address dictionary tree to obtain address keywords.

[0195] The intermediate standardized address generation module 1408 is used to extract address keywords based on the first granularity multi-level address dictionary tree to obtain target keywords, and generate intermediate standardized addresses based on the target keywords.

[0196] The target standardized address determination module 1410 is used to perform address link analysis on the intermediate standardized address based on the first granularity multi-level address dictionary tree to determine the target standardized address.

[0197] In one embodiment, the address data processing device further includes an address alias adding module 1412, which is used to obtain the alias corresponding to each address in the original multi-level standardized address library and the attribute corresponding to the alias; and add each of the aliases and the attribute corresponding to the alias to the original multi-level standardized address library to obtain the multi-level standardized address dictionary library.

[0198] In one embodiment, the dictionary tree generation module 1402 is further configured to: obtain address information corresponding to each level of address in the multi-level standardized dictionary library; store the address information in the corresponding address blocks; generate an address block dictionary based on each address block; create the root node of the first granularity multi-level address dictionary tree; obtain the address name corresponding to each address information in the address block dictionary; based on the level corresponding to the address name corresponding to each address information, create nodes corresponding to the level layer by layer starting from the root node, each node containing the address name corresponding to the level, until a corresponding node has been created for each address name corresponding to each address information, thereby generating the first granularity multi-level address dictionary tree; obtain the address name and address index number corresponding to each address information based on the address block dictionary; generate an inverted index dictionary based on the address name and address index number; and generate a second granularity multi-level address dictionary tree based on the inverted index dictionary, wherein each node of the second granularity multi-level address dictionary tree contains a character corresponding to the address name in the inverted index dictionary.

[0199] In one embodiment, the address data processing device further includes a raw address processing module 1414, which is used to obtain the raw address to be processed, remove the first special character from the raw address to be processed to obtain a first address to be processed; calculate the length of each string in the first address to be processed, delete the strings that do not meet the preset length to obtain a second address to be processed; replace the second special character corresponding to each string in the second address to be processed with a preset character to obtain a third address to be processed; and perform word segmentation on the third address to obtain a target address to be processed.

[0200] In one embodiment, the target recall address determination module 1404 is further configured to start from the starting character corresponding to the target address to be processed, and perform a search and matching on the target address to be processed based on the second granularity multi-level address dictionary tree to obtain an initial recall address. The initial recall address includes a first recall address and a second recall address, wherein the character length of the first recall address is greater than the character length of the second recall address. When the second recall address exists in the first recall address, the second recall address is deleted to obtain an intermediate recall address. The addresses obtained by incorrect matching in the intermediate recall address are deleted to obtain the target recall address.

[0201] In one embodiment, the target standardized address determination module 1410 is further configured to perform bottom-up search matching on the intermediate standardized addresses based on the first granularity multi-level address dictionary tree, calculate the number of matching address names matched by each intermediate standardized address, and use the number of matching as the matching length corresponding to the intermediate standardized address; based on each matching length and the criteria, select a second standardized address from the intermediate standardized addresses, supplement the missing address information of the second standardized address, and assign a corresponding confidence label to the second standardized address; and obtain the target standardized address from the second standardized address based on the confidence label.

[0202] In one embodiment, the address data processing device further includes a confidence labeling module 1416, used to obtain a threshold; based on each matching length, select the intermediate standardized address with the largest matching length from the intermediate standardized addresses as the third standardized address, calculate the number of address entries of the third standardized address, and use the number of address entries as the first number; when the condition is that the first number is equal to the threshold, use the third standardized address as the second standardized address, supplement the missing address information of the second standardized address, and label the second standardized address with a first confidence label; calculate the number of address entries with high-level address information in the third standardized address, and use the number of address entries as the second number; when the condition is that the first number is greater than the threshold and the second number is equal to the threshold, use the third standardized address with high-level address information as the second standardized address, supplement the missing address information of the second standardized address, and label the second standardized address with a first confidence label; calculate the number of address entries with high-level address information in the third standardized address, and use the number of address entries as the second number; when the condition is that the first number is greater than the threshold and the second number is equal to the threshold, use the third standardized address with high-level address information as the second standardized address, supplement the missing address information of the second standardized address, and label the second standardized address with a first confidence label; Standardize the address and label it with a first credibility tag; obtain addresses with the same level of suffix address information from the third standardized addresses, select the address with the latest suffix address information as the latest address, calculate the number of addresses of the latest address, and use the number of addresses of the latest address as the third number; when the first number and the second number are both greater than a threshold, and the third number is equal to the threshold, use the latest address as the second standardized address, supplement the missing address information of the second standardized address, and label the second standardized address with a second credibility tag; when the first number, the second number, and the third number are all greater than a threshold, randomly select an address from the third standardized addresses as the second standardized address, supplement the missing address information of the second standardized address, and label the second standardized address with a third credibility tag, wherein the first credibility tag is greater than the second credibility tag, and the second credibility tag is greater than the third credibility tag.

[0203] Each module in the aforementioned address data processing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0204] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 15As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores address data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When executed by the processor, the computer program implements an address data processing method.

[0205] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 16 As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements an address data processing method. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0206] Those skilled in the art will understand that Figure 15 and Figure 16The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0207] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method examples.

[0208] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0209] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps in the above method embodiments.

[0210] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data shall comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0211] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0212] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0213] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An address data processing method, characterized in that, The method includes: Obtain a multi-level standardized address dictionary, which is used to store standardized addresses; Based on the multi-level standardized address dictionary library, a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree are constructed and generated. The second-granularity multi-level address dictionary tree has a higher precision than the first-granularity multi-level address dictionary tree. Each node in the first-granularity multi-level address dictionary tree contains an address name corresponding to the level, and each node in the second-granularity multi-level address dictionary tree contains a single character corresponding to the address name. The process involves obtaining a target address to be processed, filtering the target address based on the second granularity multi-level address dictionary, and determining a target recall address. The filtering includes: starting from the first character of the target address, searching for a match in the second granularity multi-level address dictionary to obtain an initial recall address; the initial recall address includes a first recall address and a second recall address, where the character length of the first recall address is greater than the character length of the second recall address; when the second recall address exists within the first recall address, the second recall address is deleted to obtain an intermediate recall address; addresses obtained by incorrect matching in the intermediate recall address are deleted to obtain the target recall address. Based on the first granular multi-level address dictionary tree, the target recall address is segmented into keywords to obtain address keywords; Address keywords are extracted based on the first granularity multi-level address dictionary tree to obtain target keywords, and intermediate standardized addresses are generated based on the target keywords; Address link analysis is performed on the intermediate standardized addresses based on the first granularity multi-level address dictionary tree to determine the target standardized address; wherein, the address link analysis includes: performing bottom-up search matching on the intermediate standardized addresses based on the first granularity multi-level address dictionary tree, calculating the number of matching address names matched by each intermediate standardized address, and using the number of matching as the matching length corresponding to the intermediate standardized address; based on each matching length and the criteria, selecting a second standardized address from the intermediate standardized addresses, supplementing the missing address information of the second standardized address, and assigning a corresponding confidence label to the second standardized address; based on the confidence label, obtaining the target standardized address from the second standardized addresses.

2. The method according to claim 1, characterized in that, Before obtaining the multi-level standardized address dictionary, which is used to store standardized addresses, the process further includes: Obtain the aliases and attributes corresponding to each address in the original multi-level standardized address library; Each alias and its corresponding attribute are added to the original multi-level standardized address library to obtain the multi-level standardized address dictionary library.

3. The method according to claim 1, characterized in that, Based on the multi-level standardized address dictionary library, a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree are constructed, wherein the precision of the second granularity is greater than that of the first granularity, including: Obtain the address information corresponding to each level of address in the multi-level standardized address dictionary library, store the address information in the corresponding address block, and generate an address block dictionary based on each address block; Create the root node of the first-granularity multi-level address dictionary tree, obtain the address name corresponding to each address information in the address block dictionary, and based on the level corresponding to the address name corresponding to each address information, create nodes corresponding to the level layer by layer starting from the root node. Each node contains the address name corresponding to the level, until a corresponding node has been created for each address name corresponding to each address information, thus generating the first-granularity multi-level address dictionary tree. Based on the address block dictionary, obtain the address name and address index number corresponding to each address information, and generate an inverted index dictionary based on the address name and address index number; The second granularity multi-level address dictionary tree is generated based on the inverted index dictionary.

4. The method according to claim 3, characterized in that, The process involves creating nodes corresponding to different levels, starting from the root node. Each node contains an address name corresponding to its level, including: The levels are divided according to the order of province, city, district, street, road and building, with province being the highest level and building being the lowest level. Each level of the first granularity multi-level address dictionary tree corresponds to the address name of a first-level administrative division. Add the address index number corresponding to the address name to the corresponding address name to distinguish addresses with the same name in different cities.

5. The method according to claim 1, characterized in that, Before obtaining the target address to be processed, the process also includes: Obtain the original address to be processed, remove the first special character from the original address to be processed, and obtain the first address to be processed. Calculate the length of each string in the first address to be processed, delete the strings that do not meet the preset length, and obtain the second address to be processed; Replace the second special character corresponding to each string in the second address to be processed with a preset character to obtain the third address to be processed; The third address to be processed is segmented to obtain the target address to be processed.

6. The method according to claim 1, characterized in that, The step of extracting address keywords based on the first granularity multi-level address dictionary tree to obtain target keywords, and generating intermediate standardized addresses based on target keywords, includes: Referring to the address names corresponding to each node in the first granularity multi-level address dictionary tree, words that are consistent with or similar to the address names are selected from the address keywords and used as target keywords; The target keywords are combined based on the association information between them to obtain a standard multi-level address, where the association information is the distance between the target keywords. When the address information for a relevant level is missing, the address for that level is set to empty; From the standard multi-level addresses, select addresses that conform to the order of province, city, district, street, road, and building as the intermediate standardized addresses.

7. The method according to claim 1, characterized in that, The step of selecting a second standardized address from the intermediate standardized addresses based on the matching lengths and criteria, supplementing the missing address information of the second standardized address, and assigning a corresponding confidence label to the second standardized address includes: Get the threshold; Based on each of the aforementioned matching lengths, the intermediate standardized address with the largest matching length is selected from the intermediate standardized addresses as the third standardized address, and the number of address entries for the third standardized address is calculated, with the number of address entries being used as the first number. When the condition is that the first number equals the threshold, the third standardized address is used as the second standardized address to supplement the missing address information of the second standardized address, and the second standardized address is labeled with a first confidence label. Calculate the number of address entries with high-level address information in the third standardized address, and use the number of address entries as the second number; When the first number is greater than the threshold and the second number is equal to the threshold, the third standardized address with high-level address information is used as the second standardized address to supplement the missing address information of the second standardized address, and the second standardized address is labeled with a first confidence label. Obtain addresses with the same level of suffix address information from the third standardized address, select the address with the latest suffix address information as the latest address, calculate the number of addresses of the latest address, and use the number of addresses of the latest address as the third number. When the first number and the second number are both greater than the threshold, and the third number is equal to the threshold, the last address is taken as the second standardized address, the missing address information of the second standardized address is supplemented, and the second standardized address is labeled with a second confidence label. When the first number, the second number, and the third number are all greater than the threshold, an address is randomly selected from the third standardized address as the second standardized address to supplement the missing address information of the second standardized address, and a third confidence label is marked on the second standardized address. The first confidence label is greater than the second confidence label, and the second confidence label is greater than the third confidence label.

8. An address data processing device, characterized in that, The device includes: A multi-level standardized address dictionary acquisition module is used to acquire a multi-level standardized address dictionary, which is used to store standardized addresses; The dictionary tree generation module is used to construct and generate a first-granularity multi-level address dictionary tree and a second-granularity multi-level address dictionary tree based on the multi-level standardized address dictionary library. The second-granularity multi-level address dictionary tree has a higher precision than the first-granularity multi-level address dictionary tree. Each node in the first-granularity multi-level address dictionary tree contains an address name corresponding to the level, and each node in the second-granularity multi-level address dictionary tree contains a single character corresponding to the address name. The target recall address determination module is used to obtain the target address to be processed, filter the target address to be processed based on the second granularity multi-level address dictionary tree, and determine the target recall address; wherein, the target recall address determination module is further specifically used to: starting from the starting character of the target address to be processed, search for a match in the second granularity multi-level address dictionary tree to obtain an initial recall address; the initial recall address includes a first recall address and a second recall address, the character length of the first recall address is greater than the character length of the second recall address; when the second recall address exists in the first recall address, the second recall address is deleted to obtain an intermediate recall address; the addresses obtained by incorrect matching in the intermediate recall address are deleted to obtain the target recall address; The address keyword determination module is used to perform keyword segmentation on the target recall address based on the first granular multi-level address dictionary tree to obtain address keywords; The intermediate standardized address generation module is used to extract address keywords based on the first granularity multi-level address dictionary tree to obtain target keywords, and generate intermediate standardized addresses based on the target keywords; The target standardized address determination module is used to search and match the intermediate standardized addresses based on the first granularity multi-level address dictionary tree to perform address link analysis from low-level addresses to high-level addresses and determine the target standardized address. Specifically, the intermediate standardized address generation module is further used to perform bottom-up search and matching on the intermediate standardized addresses based on the first granularity multi-level address dictionary tree, calculate the number of match names for each intermediate standardized address, and use the number of match names as the match length corresponding to the intermediate standardized address. Based on each match length and a condition, a second standardized address is selected from the intermediate standardized addresses, missing address information is supplemented, and a corresponding confidence label is assigned to the second standardized address. Based on the confidence label, the target standardized address is obtained from the second standardized addresses.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images