House property data intelligent matching and real estate table consistency management method and system

Abstract

This application discloses a method and system for intelligent matching of real estate data and consistency governance of building lists, belonging to the field of real estate data processing technology. It includes pre-matching building and house data based on the data characteristics of each type of data source, generating corresponding data linking logic rules, performing linking operations on related building and house data, establishing correspondences between building and house data from different data sources, establishing related data pairs, establishing a one-to-one mapping relationship between houses and buildings, designing subject identification rules, extracting subject information and related ownership information from the mapping relationship, monitoring the synchronization and consistency of subject information and related ownership information during online signing or management, issuing warnings when data deviations or synchronization failures are detected, locating abnormal nodes in the mapping relationship, and completing error correction based on preset rules. This improves matching accuracy, increases the proportion of automatic linking, and reduces the linking error rate.

Description

Technical Field

[0001] This application belongs to the field of real estate data processing technology, specifically, it relates to a method and system for intelligent matching of real estate data and consistency management of building information tables. Background Technology

[0002] In today's booming real estate market, the management and utilization of real estate data is of paramount importance to numerous stakeholders, including real estate management departments, financial institutions, and real estate companies. Accurate, comprehensive, and consistent data provides strong support for market analysis, policy formulation, strategic planning, and business decision-making. It offers clear and accurate market dynamics information, helping all parties grasp market trends and make informed and rational decisions.

[0003] The current state of data management in the real estate sector is far from ideal. Multiple key business systems, such as the maintenance fund management system, online contract signing system, and real estate system, operate independently. Over the long term, each system has built and accumulated data based on its own business needs and technical architecture. Due to differences in construction timelines and varying technical standards adopted at different periods, coupled with a lack of unified standards and constraints for data entry, data on buildings and properties across these systems exhibits inconsistent formats and information, creating a chaotic situation. This fragmented and inconsistent data severely hinders data sharing and utilization efficiency. In actual business operations, systems often need to collaborate to complete complex processes, but the inconsistency and fragmentation of data lead to poor information transmission. This not only increases the complexity and time cost of business processing but also easily results in data errors and duplicate entries, ultimately affecting the accuracy and efficiency of business operations and failing to meet the demands of efficient and precise development in the real estate business. Summary of the Invention

[0004] To address the aforementioned problems and technical deficiencies, this application adopts the following technical solution: a method for intelligent matching of real estate data and consistency governance of building information lists, comprising the following steps:

[0005] Building and house data are obtained from different types of data sources, and a preset matching algorithm is used to pre-match the building and house data according to the data characteristics of each type of data source;

[0006] Based on the pre-matching results of buildings and houses, corresponding data attachment logic rules are generated. The data attachment operation is performed on the associated building and house data according to the data attachment logic rules to establish the correspondence between building and house data in different data sources and to establish data pairs with related relationships.

[0007] Based on the data, a one-to-one mapping relationship between houses and buildings is established, subject identification rules are designed, and subject information and related ownership information corresponding to the subject information are extracted from the mapping relationship.

[0008] During online contract signing or management, the system monitors the synchronization and consistency of the main entity information and related ownership information. When data deviation or synchronization failure is detected, an early warning is issued, abnormal nodes in the mapping relationship are located, and error correction is completed based on preset rules.

[0009] Preferably, the pre-matching of building and housing data includes:

[0010] The building names and house names are segmented into words to generate corresponding word vector sets;

[0011] The similarity between building names is calculated based on the word vector set. Based on the similarity calculation results, building pairs with similarity higher than a preset threshold are selected as candidate building matching pairs.

[0012] Verify the geospatial correlation and development entity of the candidate building matching pairs;

[0013] Based on the word vector set, the name similarity of the houses in the verified candidate building matching pairs is calculated, and the house pairs with similarity higher than the preset threshold are selected as candidate house matching pairs.

[0014] Verify the owner information and property attribute information of the candidate house matching pairs;

[0015] Confirm the association between successfully verified buildings and houses, and mark the reasons for unmatched data.

[0016] Preferably, the data attachment logic rules include:

[0017] The high-confidence attachment rule automatically attaches data when the pre-match similarity is greater than the first threshold. It takes the coding and ownership information of real estate registration data as the main data, integrates the filing status of housing and construction filing data and the update information of third-party data, and generates an association structure of main data and supplementary data after attachment. At the same time, it records the mapping relationship of the attached fields.

[0018] The medium-confidence attachment rule performs semi-automatic attachment when the first threshold > pre-match similarity ≥ the second threshold. For quantifiable deviations such as area deviation and encoding format differences, it automatically generates deviation correction formulas and automatically attaches after correction. For address synonym differences, it performs manual verification. After manual confirmation, the attachment is completed, and the manual confirmation results are fed back to the rule base to optimize subsequent attachment rules.

[0019] For low-confidence data linking rules, when the second threshold is greater than the pre-match similarity, manual guidance is provided for linking. The system marks the unmatched information fields, generates linking guidance prompts, and manual field mapping is performed. The system records this mapping relationship as a template rule for subsequent linking of similar data.

[0020] After each connection is completed, the connection success rate and the number of abnormal connections are counted. Based on the statistical results, the weight allocation, first threshold and second threshold of the data connection logic rules are adjusted.

[0021] Furthermore, the association hierarchy is divided into three levels, including:

[0022] The core information of the building is used as the association field to match the building data from different data sources one by one, and the association source is marked to generate a first-level building association.

[0023] By using core housing information as the associated field, housing data from different data sources are mapped one-to-one and associated with the corresponding building relationships, generating secondary housing associations.

[0024] The property ownership information is linked with the registration information and third-party updated information to form a complete association chain, generating a three-level ownership association.

[0025] Each level of the relationship contains traceability information. When data anomalies occur, the connection steps can be traced back to pinpoint the cause of the anomaly.

[0026] Furthermore, the data pairs are two types of dedicated data pairs.

[0027] The first category is benchmark and non-benchmark data pairs, labeled with matching similarity and attachment rules;

[0028] The second type is associated data pairs, which are labeled with association levels and traceability information. The two types of data pairs are related to each other to generate complete associated data.

[0029] Furthermore, the mapping relationship is established by first establishing mapping constraint rules, then using a tree structure to store the mapping relationship. The root node of the tree structure is used as the building complex, the child nodes as the building blocks, and the leaf nodes as the houses. Each node is associated with corresponding data pairs and traceability information. A hash index is established based on the building block code and the house code to realize the query of the mapping relationship.

[0030] Furthermore, the mapping constraint rules include:

[0031] A building can only correspond to a unit within its own building complex. By locking the building complex area, cross-building mapping is prohibited.

[0032] One house can only correspond to one building, and the validity of the association is verified by the house code;

[0033] The mapping relationship must be unique. The primary key is combined with the building code and the house code to avoid duplicate mappings.

[0034] Furthermore, the subject identification rule is established based on predefined building subject information and house subject information, including:

[0035] Building entity identification rules: Based on the building data pairs in the mapping relationship, extract the building code and address fields common to each data source, standardize the address through natural language processing technology, and generate a unique identifier for the building entity by cross-validation with building parameters; if there are parameter conflicts, the parameters of the real estate registration data source shall be used first.

[0036] Building entity identification rules: Based on the building data pairs in the mapping relationship, extract the building code and area fields common to each data source, correct the area deviation, and combine the unit type cross-validation to generate a unique identifier for the building entity. At the same time, associate it with the corresponding unique identifier for the building entity to make the building and the building correspond one-to-one.

[0037] Rules for extracting ownership-related information: Extract core ownership information from real estate registration data sources, extract ownership association information from supplementary data sources, and extract ownership feature information using field association and semantic extraction; for ownership association information in unstructured text, supplement information using NLP semantic extraction technology.

[0038] Furthermore, the abnormal node location is based on a tree structure of mapping relationships and hash indexes to locate the specific location, unique identifier, abnormal type, data source and specific fields involved in the abnormality, occurrence time, and associated connection rules and mapping relationships of the abnormal node.

[0039] The real estate data intelligent matching and building list consistency governance system includes:

[0040] The pre-matching module is used to obtain building and house data from different types of data sources. It uses a preset matching algorithm to pre-match the building and house data according to the data characteristics of each type of data source.

[0041] The data pair module generates corresponding data attachment logic rules based on the pre-matching results of buildings and houses. It then performs attachment operations on the associated building and house data according to the data attachment logic rules, establishes the correspondence between building and house data in different data sources, and creates data pairs with related relationships.

[0042] The information extraction module establishes a one-to-one mapping relationship between houses and buildings based on data pairs, designs subject identification rules, and extracts subject information and related ownership information corresponding to the subject information from the mapping relationship.

[0043] The anomaly monitoring module monitors the synchronization and consistency of the main entity information and related ownership information during online signing or management. When data deviation or synchronization failure is detected, it issues an early warning, locates the abnormal node in the mapping relationship, and completes error correction based on preset rules.

[0044] Compared to existing technologies, the beneficial effects of this application are as follows:

[0045] (1) This application designs a differentiated matching logic, weight allocation and deviation correction scheme, and incorporates exclusive processing such as property code standardization, address synonym matching and area deviation correction to improve the matching accuracy;

[0046] (2) This application designs a dynamic self-optimizing real estate data linking rule and a multi-level traceable association mechanism, which is different from the fixed linking rule of the existing technology. The linking rule is dynamically generated based on the pre-matching result, and the linking is graded according to the confidence level. The rule can be automatically optimized according to the linking effect, and a three-level association relationship and blockchain evidence storage and traceability mechanism are established to improve the automatic linking ratio and reduce the linking error rate.

[0047] (3) This application establishes a one-to-one mapping constraint between buildings and houses for real estate and logic for subject identification and ownership extraction. It is different from general mapping and subject identification technologies. It sets three types of exclusive constraint rules: building scope, building binding, and mapping uniqueness. It deeply binds subject identification and ownership extraction and adopts a deduplication strategy that combines authoritative data sources and supplementary data sources to accurately identify subjects. Attached Figure Description

[0048] In the attached diagram:

[0049] Figure 1 This is a schematic diagram of the method steps in an embodiment of this application;

[0050] Figure 2 This is a schematic diagram of the system structure according to an embodiment of this application. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of this application, but not all embodiments. Generally, the components of the embodiments of this application described and shown in the accompanying drawings can be arranged and designed in various different configurations.

[0052] Example 1

[0053] like Figure 1 As shown, the intelligent matching of real estate data and the method for maintaining consistency with building lists include the following steps:

[0054] Building and house data are obtained from different types of data sources, and a preset matching algorithm is used to pre-match the building and house data according to the data characteristics of each type of data source;

[0055] Pre-matching of building and housing data includes:

[0056] The building names and house names are segmented into words to generate corresponding word vector sets;

[0057] The similarity between building names is calculated based on the word vector set. Based on the similarity calculation results, building pairs with similarity higher than a preset threshold are selected as candidate building matching pairs.

[0058] Verify the geospatial correlation and development entity of the candidate building matching pairs;

[0059] Based on the word vector set, the name similarity of the houses in the verified candidate building matching pairs is calculated, and the house pairs with similarity higher than the preset threshold are selected as candidate house matching pairs.

[0060] Verify the owner information and property attribute information of the candidate house matching pairs;

[0061] Confirm the association between successfully verified buildings and houses, and mark the reasons for unmatched data.

[0062] Based on the pre-matching results of buildings and houses, corresponding data attachment logic rules are generated. The data attachment operation is performed on the associated building and house data according to the data attachment logic rules to establish the correspondence between building and house data in different data sources and to establish data pairs with related relationships.

[0063] The data connection logic rules include:

[0064] The high-confidence attachment rule automatically attaches data when the pre-match similarity is greater than the first threshold. It takes the coding and ownership information of real estate registration data as the main data, integrates the filing status of housing and construction filing data and the update information of third-party data, and generates an association structure of main data and supplementary data after attachment. At the same time, it records the mapping relationship of the attached fields.

[0065] The medium-confidence attachment rule performs semi-automatic attachment when the first threshold > pre-match similarity ≥ the second threshold. For quantifiable deviations such as area deviation and encoding format differences, it automatically generates deviation correction formulas and automatically attaches after correction. For address synonym differences, it performs manual verification. After manual confirmation, the attachment is completed, and the manual confirmation results are fed back to the rule base to optimize subsequent attachment rules.

[0066] For low-confidence data linking rules, when the second threshold is greater than the pre-match similarity, manual guidance is provided for linking. The system marks the unmatched information fields, generates linking guidance prompts, and manual field mapping is performed. The system records this mapping relationship as a template rule for subsequent linking of similar data.

[0067] After each connection is completed, the connection success rate and the number of abnormal connections are counted. Based on the statistical results, the weight allocation, first threshold and second threshold of the data connection logic rules are adjusted.

[0068] The hierarchy of relationships is divided into three levels, including:

[0069] The core information of the building is used as the association field to match the building data from different data sources one by one, and the association source is marked to generate a first-level building association.

[0070] By using core housing information as the associated field, housing data from different data sources are mapped one-to-one and associated with the corresponding building relationships, generating secondary housing associations.

[0071] The property ownership information is linked with the registration information and third-party updated information to form a complete association chain, generating a three-level ownership association.

[0072] Each level of the relationship contains traceability information. When data anomalies occur, the connection steps can be traced back to pinpoint the cause of the anomaly.

[0073] The data pairs are divided into two categories of specific data pairs.

[0074] The first category is benchmark and non-benchmark data pairs, labeled with matching similarity and attachment rules;

[0075] The second type is associated data pairs, which are labeled with association levels and traceability information. The two types of data pairs are related to each other to generate complete associated data.

[0076] Based on the data, a one-to-one mapping relationship between houses and buildings is established, subject identification rules are designed, and subject information and related ownership information corresponding to the subject information are extracted from the mapping relationship.

[0077] The mapping relationship is established by first establishing mapping constraint rules, then using a tree structure to store the mapping relationship. The root node of the tree structure is used as the building, the child nodes are used as the building blocks, and the leaf nodes are used as the houses. Each node is associated with a corresponding data pair and traceability information. A hash index is built based on the building code and the house code to realize the query of the mapping relationship.

[0078] Mapping constraint rules include:

[0079] A building can only correspond to a unit within its own building complex. By locking the building complex area, cross-building mapping is prohibited.

[0080] One house can only correspond to one building, and the validity of the association is verified by the house code;

[0081] The mapping relationship must be unique. The primary key is combined with the building code and the house code to avoid duplicate mappings.

[0082] The entity identification rules are established based on predefined building entity information and house entity information, including:

[0083] Building entity identification rules: Based on the building data pairs in the mapping relationship, extract the building code and address fields common to each data source, standardize the address through natural language processing technology, and generate a unique identifier for the building entity by cross-validation with building parameters; if there are parameter conflicts, the parameters of the real estate registration data source shall be used first.

[0084] Building entity identification rules: Based on the building data pairs in the mapping relationship, extract the building code and area fields common to each data source, correct the area deviation, and combine the unit type cross-validation to generate a unique identifier for the building entity. At the same time, associate it with the corresponding unique identifier for the building entity to make the building and the building correspond one-to-one.

[0085] Rules for extracting ownership-related information: Extract core ownership information from real estate registration data sources, extract ownership association information from supplementary data sources, and extract ownership feature information using field association and semantic extraction; for ownership association information in unstructured text, supplement information using NLP semantic extraction technology.

[0086] During online contract signing or management, the system monitors the synchronization and consistency of the main entity information and related ownership information. When data deviation or synchronization failure is detected, an early warning is issued, abnormal nodes in the mapping relationship are located, and error correction is completed based on preset rules.

[0087] Abnormal node location is based on a tree structure and hash index of mapping relationships to locate the specific location, unique identifier, abnormal type, data source and specific fields involved in the abnormality, occurrence time, and associated rules and mapping relationships of the abnormality.

[0088] Based on this, it should be noted that government and public service systems related to real estate and property owners can rely on this system solution to perform data integration, governance, and push, so as to achieve efficient data integration, accurate governance, and real-time push.

[0089] Specific application examples include, but are not limited to, the following aspects:

[0090] First, taking the "Pre-sale Funds Supervision System" as an example, this system, as an important component of real estate supervision, has a close business relationship with the online contract signing system. With the support of this system solution, it can realize the automatic push and synchronous update of key data based on the deep connection between the pre-sale funds supervision system and the online contract signing system. This data includes, but is not limited to, the addition, modification and deletion of building and house information, the establishment, modification and termination status of supervision business, the signing, modification and cancellation of subscription agreements, the signing, filing, modification and cancellation of contracts, as well as key changes of enterprises and development projects. Through this mechanism, the timeliness and accuracy of pre-sale funds supervision are ensured, providing strong support for government decision-making.

[0091] Secondly, taking the "natural gas system" as an example, this system, as an infrastructure for residential services, stores a large amount of user information. These users may be property owners or tenants. Through this system solution, based on the connection between the maintenance fund management system and the natural gas system, the actual owner information of the property can be effectively identified and obtained. This process not only improves the efficiency and accuracy of data acquisition, but also provides valuable data support for subsequent property management, service optimization and policy formulation.

[0092] Example 2

[0093] like Figure 2 As shown, the real estate data intelligent matching and building list consistency governance system includes:

[0094] The pre-matching module is used to obtain building and house data from different types of data sources. It uses a preset matching algorithm to pre-match the building and house data according to the data characteristics of each type of data source.

[0095] The data pair module generates corresponding data attachment logic rules based on the pre-matching results of buildings and houses. It then performs attachment operations on the associated building and house data according to the data attachment logic rules, establishes the correspondence between building and house data in different data sources, and creates data pairs with related relationships.

[0096] The information extraction module establishes a one-to-one mapping relationship between houses and buildings based on data pairs, designs subject identification rules, and extracts subject information and related ownership information corresponding to the subject information from the mapping relationship.

[0097] The anomaly monitoring module monitors the synchronization and consistency of the main entity information and related ownership information during online signing or management. When data deviation or synchronization failure is detected, it issues an early warning, locates the abnormal node in the mapping relationship, and completes error correction based on preset rules.

[0098] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications, improvements, and substitutions without departing from the concept of this application, and these all fall within the protection scope of this application.

Claims

1. A method for intelligent matching of real estate data and consistency management of building listings, characterized in that: Includes the following steps: Building and house data are obtained from different types of data sources, and a preset matching algorithm is used to pre-match the building and house data according to the data characteristics of each type of data source; Based on the pre-matching results of buildings and houses, corresponding data attachment logic rules are generated. The data attachment operation is performed on the associated building and house data according to the data attachment logic rules to establish the correspondence between building and house data in different data sources and to establish data pairs with related relationships. The data linking logic rules include: The high-confidence attachment rule automatically attaches data when the pre-match similarity is greater than the first threshold. It takes the coding and ownership information of real estate registration data as the main data, integrates the filing status of housing and construction filing data and the update information of third-party data, and generates an association structure of main data and supplementary data after attachment. At the same time, it records the mapping relationship of the attached fields. The medium-confidence attachment rule performs semi-automatic attachment when the first threshold > pre-match similarity ≥ the second threshold. For quantifiable deviations such as area deviation and encoding format differences, it automatically generates deviation correction formulas and automatically attaches after correction. For address synonym differences, it performs manual verification. After manual confirmation, the attachment is completed, and the manual confirmation results are fed back to the rule base to optimize subsequent attachment rules. For low-confidence data linking rules, when the second threshold is greater than the pre-match similarity, manual guidance is provided for linking. The system marks the unmatched information fields, generates linking guidance prompts, and manual field mapping is performed. The system records this mapping relationship as a template rule for subsequent linking of similar data. After each connection is completed, the connection success rate and the number of abnormal connections are counted. Based on the statistical results, the weight allocation, first threshold and second threshold of the data connection logic rules are adjusted. Based on the data, a one-to-one mapping relationship between houses and buildings is established, subject identification rules are designed, and subject information and related ownership information corresponding to the subject information are extracted from the mapping relationship. The entity identification rules are established based on predefined building entity information and house entity information, including: Building entity identification rules: Based on the building data pairs in the mapping relationship, extract the building code and address fields common to each data source, standardize the address through natural language processing technology, and generate a unique identifier for the building entity by cross-validation with building parameters; if there are parameter conflicts, the parameters of the real estate registration data source shall be used first. Building entity identification rules: Based on the building data pairs in the mapping relationship, extract the building code and area fields common to each data source, correct the area deviation, and combine the unit type cross-validation to generate a unique identifier for the building entity. At the same time, associate it with the corresponding unique identifier for the building entity to make the building and the building correspond one-to-one. Rules for extracting ownership-related information: Extract core ownership information from real estate registration data sources, extract ownership-related information from supplementary data sources, and extract ownership feature information using field association and semantic extraction; for ownership-related information in unstructured text, supplement information using NLP semantic extraction technology. During online contract signing or management, the system monitors the synchronization and consistency of the main entity information and related ownership information. When data deviation or synchronization failure is detected, an early warning is issued, abnormal nodes in the mapping relationship are located, and error correction is completed based on preset rules.

2. The method for intelligent matching of real estate data and consistency management of building lists according to claim 1, characterized in that, The pre-matching of building and housing data includes: The building names and house names are segmented into words to generate corresponding word vector sets; The similarity between building names is calculated based on the word vector set. Based on the similarity calculation results, building pairs with similarity higher than a preset threshold are selected as candidate building matching pairs. Verify the geospatial correlation and development entity of the candidate building matching pairs; Based on the word vector set, the name similarity of the houses in the verified candidate building matching pairs is calculated, and the house pairs with similarity higher than the preset threshold are selected as candidate house matching pairs. Verify the owner information and property attribute information of the candidate house matching pairs; Confirm the association between successfully verified buildings and houses, and mark the reasons for unmatched data.

3. The method for intelligent matching of real estate data and consistency management of building lists according to claim 1, characterized in that, The association hierarchy is divided into three levels, including: The core information of the building is used as the association field to match the building data from different data sources one by one, and the association source is marked to generate a first-level building association. By using core housing information as the associated field, housing data from different data sources are mapped one-to-one and associated with the corresponding building relationships, generating secondary housing associations. The property ownership information is linked with the registration information and third-party updated information to form a complete association chain, generating a three-level ownership association. Each level of the relationship contains traceability information. When data anomalies occur, the connection steps can be traced back to pinpoint the cause of the anomaly.

4. The method for intelligent matching of real estate data and consistency management of building lists according to claim 3, characterized in that, The data pairs are two types of dedicated data pairs. The first category is benchmark and non-benchmark data pairs, labeled with matching similarity and attachment rules; The second type is associated data pairs, which are labeled with association levels and traceability information. The two types of data pairs are related to each other to generate complete associated data.

5. The method for intelligent matching of real estate data and consistency management of building lists according to claim 3, characterized in that, The mapping relationship is established by first establishing mapping constraint rules, then using a tree structure to store the mapping relationship. The root node of the tree structure is used as the building, the child nodes are used as the building blocks, and the leaf nodes are used as the houses. Each node is associated with a corresponding data pair and traceability information. A hash index is established based on the building code and the house code to realize the query of the mapping relationship.

6. The method for intelligent matching of real estate data and consistency management of building lists according to claim 5, characterized in that, The mapping constraint rules include: A building can only correspond to a unit within its own building complex. By locking the building complex area, cross-building mapping is prohibited. One house can only correspond to one building, and the validity of the association is verified by the house code; The mapping relationship must be unique. The primary key is combined with the building code and the house code to avoid duplicate mappings.

7. The method for intelligent matching of real estate data and consistency management of building lists according to claim 5, characterized in that, The abnormal node location is based on a tree structure and hash index of mapping relationships to locate the specific location, unique identifier, abnormal type, data source and specific fields involved in the abnormality, occurrence time, and associated connection rules and mapping relationships of the abnormal node.

8. A real estate data intelligent matching and building list consistency management system, characterized in that: include: The pre-matching module is used to obtain building and house data from different types of data sources. It uses a preset matching algorithm to pre-match the building and house data according to the data characteristics of each type of data source. The data pair module generates corresponding data attachment logic rules based on the pre-matching results of buildings and houses. It then performs attachment operations on the associated building and house data according to the data attachment logic rules, establishes the correspondence between building and house data in different data sources, and creates data pairs with related relationships. The data linking logic rules include: The high-confidence attachment rule automatically attaches data when the pre-match similarity is greater than the first threshold. It takes the coding and ownership information of real estate registration data as the main data, integrates the filing status of housing and construction filing data and the update information of third-party data, and generates an association structure of main data and supplementary data after attachment. At the same time, it records the mapping relationship of the attached fields. The medium-confidence attachment rule performs semi-automatic attachment when the first threshold > pre-match similarity ≥ the second threshold. For quantifiable deviations such as area deviation and encoding format differences, it automatically generates deviation correction formulas and automatically attaches after correction. For address synonym differences, it performs manual verification. After manual confirmation, the attachment is completed, and the manual confirmation results are fed back to the rule base to optimize subsequent attachment rules. For low-confidence data linking rules, when the second threshold is greater than the pre-match similarity, manual guidance is provided for linking. The system marks the unmatched information fields, generates linking guidance prompts, and manual field mapping is performed. The system records this mapping relationship as a template rule for subsequent linking of similar data. After each connection is completed, the connection success rate and the number of abnormal connections are counted. Based on the statistical results, the weight allocation, first threshold and second threshold of the data connection logic rules are adjusted. The information extraction module establishes a one-to-one mapping relationship between houses and buildings based on data pairs, designs subject identification rules, and extracts subject information and related ownership information corresponding to the subject information from the mapping relationship. The entity identification rules are established based on predefined building entity information and house entity information, including: Building entity identification rules: Based on the building data pairs in the mapping relationship, extract the building code and address fields common to each data source, standardize the address through natural language processing technology, and generate a unique identifier for the building entity by cross-validation with building parameters; if there are parameter conflicts, the parameters of the real estate registration data source shall be used first. Building entity identification rules: Based on the building data pairs in the mapping relationship, extract the building code and area fields common to each data source, correct the area deviation, and combine the unit type cross-validation to generate a unique identifier for the building entity. At the same time, associate it with the corresponding unique identifier for the building entity to make the building and the building correspond one-to-one. Rules for extracting ownership-related information: Core ownership information is extracted from real estate registration data sources, and ownership association information is extracted from supplementary data sources. Field association and semantic extraction are used to extract ownership feature information. For ownership association information in unstructured text, NLP semantic extraction technology is used to supplement the information. The anomaly monitoring module monitors the synchronization and consistency of main information and related ownership information in real time. When data deviation or synchronization failure is detected, it issues an early warning, locates the abnormal node in the mapping relationship, and completes error correction based on preset rules.

Images