A housing multi-source information management method, system and storage medium
By acquiring and processing basic housing information and user interaction data, generating feature vectors and integrating them into the database, recommending housing information and predicting rental and sales trends, the problem of low efficiency in traditional housing management is solved, and efficient information management and data support are achieved.
Patent Information
- Application Number
- CN202311333988.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-16
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2043-10-16
AI Technical Summary
Traditional housing management methods are inefficient, information management is inaccurate, and they cannot meet the rapidly growing customer demand, leading to difficulties in property sales and rental management.
By acquiring basic information about the target property and business requirements, data preprocessing and business workflow extraction are performed to generate user feature vectors and property feature vectors. The property database is then used for information recommendation and evaluation to predict rental and sales information and optimize property management.
It has enabled the scientific management of housing information, improved the quality of information services, reduced management costs, provided accurate data support, and met the diverse needs of customers.
Smart Images

Figure CN117291409B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of house information management, and more specifically, to a house multi-source information management method, system and storage medium. Background Art
[0002] With the continuous development of computer technology and information construction, all walks of life have entered the stage of information and digital development. The real estate industry is also moving towards information construction. With the continuous development of real estate houses and the continuous increase in house buyers, the traditional and crude house management methods of real estate sales and leasing companies can no longer meet the needs of the rapidly growing number of buyers and leasers. For the construction, sales and leasing of houses and the subsequent property management, a large amount of information needs to be managed and maintained. However, many real estate companies are still in a completely manual and manual working state. This has brought great difficulties to the construction, sales, leasing and management of the real estate industry.
[0003] Housing resources are the foundation of real estate companies. Effective management of housing can help sales staff quickly understand the specific conditions and sales status of housing so as to make corresponding sales plans. It also provides data support for management personnel to make housing sales plans and related major decisions. Real estate agencies that mainly rely on manual registration often have piles of documents that need to be sorted manually. The collection, update, and query of housing information are very slow and prone to errors. The information management efficiency is low, and high-quality information services cannot be guaranteed. It is difficult to quickly meet customer needs. Therefore, how to strengthen the scientific and information management of customer sources, housing sources, and performance information in real estate sales is an urgent problem to be solved. Summary of the invention
[0004] In order to solve the above technical problems, the present invention proposes a housing multi-source information management method, system and storage medium.
[0005] A first aspect of the present invention provides a housing multi-source information management method, comprising:
[0006] Obtain basic information of the target house and corresponding business requirements, perform data preprocessing on the basic information, extract the business workflow corresponding to the business requirements, and use the preprocessed basic information to fill in data for the business workflow;
[0007] Obtain corresponding permissions according to the user's identity information and login information, obtain corresponding pages and house data based on the permissions, obtain the user's house interaction data, and generate a user feature vector through the user's identity information and house interaction data;
[0008] Extracting the house feature vector of the target house based on the business workflow, matching the address code, associating and integrating it into the house database, and using the house database to recommend house information to the user;
[0009] The search popularity of the target house and the user's interest are obtained for data integration, the target house is evaluated, the house rental and sale information is predicted based on the evaluation result, and the optimization direction of the target house is determined based on the predicted house rental and sale information.
[0010] In this solution, the basic information is preprocessed to extract the business workflow corresponding to the business needs, and the preprocessed basic information is used to fill the business workflow with data, specifically:
[0011] Extracting basic information of a target house, judging whether the target house is a newly added house according to the basic information, generating different information acquisition templates for the target house according to the judgment result, and collecting multi-source information of the target house or updating multi-source information according to the information acquisition template;
[0012] Clean the collected or updated multi-source information, match the business requirements of the target house, obtain the corresponding functional requirements, data requirements and non-functional requirements for the target house according to the business requirements, and generate the business workflow corresponding to the business requirements;
[0013] The multi-source information of the target house needs to be verified using the data of each node in the business workflow. If the verification fails, the multi-source information is collected again, and the business workflow is filled with the multi-source data that passes the verification;
[0014] When the target house is not a newly added house, the deviation between the updated multi-source information and the original multi-source information is used to compensate the data for the workflow node, extract the main process path corresponding to the original multi-source information, and expand the business needs through the multi-source information deviation analysis;
[0015] According to the business demand expansion, the starting point is located in the main process path, and the historical process path corresponding to the business demand and with the same predecessor task is retrieved through instances. The historical process paths are aggregated to remove redundant nodes, and the sub-process paths corresponding to the business demand expansion are obtained, and the main process path is expanded and updated.
[0016] In this solution, the corresponding page and house data are obtained based on the permissions, the user's house interaction data is obtained, and the user feature vector is generated through the user's identity information and house interaction data, specifically:
[0017] Obtaining user identity information and login information, extracting the user's login request, verification token and terminal address information based on the identity information and login information, and determining whether the user is a legitimate user based on the verification token;
[0018] If the user is a legitimate user, determine whether it is the first time for the user to visit, edit the user information and complete the identity information according to the determination result, query the user's authority level through the identity information to generate the corresponding access page and house information;
[0019] Obtaining the house interaction data of the user on the visited page, setting an interaction threshold, using the interaction threshold to filter the house interaction data, and obtaining weight information of the house interaction data according to the interaction frequency and interaction duration;
[0020] The filtered house interaction data is combined with the identity information for feature encoding, and the attention mechanism is introduced in the encoder network to set the attention weight for the house interaction data. The encoded features are fused in combination with the weight information, and the user feature vector is generated through the decoder.
[0021] In this solution, the house feature vector of the target house is extracted based on the business workflow, and the address code is matched and associated and integrated and stored in the house database. Specifically:
[0022] Acquire two-dimensional map information of the target area, grid the two-dimensional map information, encode each grid, count the house information in the grid, classify and encode the house information using business data, and generate address coding criteria for the house;
[0023] Determine the multi-source information after the target house is matched according to the business workflow, extract the location representation data in the matched multi-source information, perform format conversion and coordinate conversion on the location representation data, and generate a corresponding address code as an identification code according to the address coding rule;
[0024] The non-position representation data in the multi-source information is projected and transformed to generate the graph representation data of the multi-source information, the address code is combined with the generated graph representation data to generate the house feature vector of the target house, and the result is stored in the house database.
[0025] In this solution, the housing database is used to recommend housing information to users, specifically:
[0026] Acquire user vector features in a housing database, perform clustering on the user vector features, divide the user vector features into a plurality of clusters, calculate a silhouette coefficient of each cluster, and determine an optimal number of clusters corresponding to the user vector features according to the silhouette coefficient;
[0027] Selecting an initial cluster center according to the number of clusters, determining clustering results of different user categories according to iterative clustering and division, and obtaining average user characteristics of different user categories according to the clustering results;
[0028] Obtain a user vector feature of the current user, obtain a category label of the current user based on the comparison between the user vector feature and the average features of users of different categories, and import the user vector feature into the LSTM model to learn the temporal features of user preferences from the house interaction data;
[0029] Capturing important house interaction data in the time series features through a self-attention network, obtaining weight information of the important house interaction data, and weighting the time series features using the weight information to obtain the preference features of the current user;
[0030] Representing the data in the housing database as a graph to construct an undirected graph, learning the undirected graph through a graph convolutional network, and generating a graph structure according to a user vector feature of a current user;
[0031] Obtain similar users through the category label of the current user, generate a neighbor matrix according to the similar users and the house feature vectors corresponding to the preference features, and use the neighbor matrix to represent the graph structure;
[0032] Through the message propagation and neighbor aggregation mechanism of the graph convolutional neural network, the graph structure is learned, the user vector features and house feature vectors are updated, the inner product of the updated user vector features and house feature vectors is obtained, and house information is recommended for the current user based on the inner product.
[0033] In this solution, the search popularity of the target house and the user's interest are obtained for data integration, the target house is evaluated, the house rental and sale information is predicted based on the evaluation results, and the optimization direction of the target house is determined based on the predicted house rental and sale information, specifically:
[0034] The search popularity of the target house is obtained based on the number of views, recommendations, and promotions of the target house within a preset time, as well as the time interval between the first and last views of the user, and the weight information of the target house in the user preference features is obtained based on the user feature vector in the house database within the preset time;
[0035] The weight information of each customer in different user categories is averaged to obtain the interest of different user categories in the target house, and the interest of different user categories in the target house is corrected by using the average transaction time of the target house;
[0036] Attention weights are introduced to evaluate target houses by weighted calculation based on the search popularity of the target house and the user's interest. A house rental and sale prediction model is built based on deep learning methods and trained through house rental and sale examples.
[0037] The evaluation result of the target house is used as the input of the house rental and sale prediction model to obtain the predicted rental and sale period of the target house. When the predicted rental and sale period is greater than a preset time threshold, similar instances whose rental and sale periods meet the standard threshold are screened according to the house rental and sale instances;
[0038] An optimization direction of the target house is determined according to the deviation between the similar instance and the target house.
[0039] The second aspect of the present invention further provides a housing multi-source information management system, the system comprising: a memory, a processor, the memory comprising a housing multi-source information management method program, the housing multi-source information management method program when executed by the processor to implement the following steps:
[0040] Obtain basic information of the target house and corresponding business requirements, perform data preprocessing on the basic information, extract the business workflow corresponding to the business requirements, and use the preprocessed basic information to fill in data for the business workflow;
[0041] Obtain corresponding permissions according to the user's identity information and login information, obtain corresponding pages and house data based on the permissions, obtain the user's house interaction data, and generate a user feature vector through the user's identity information and house interaction data;
[0042] Extracting the house feature vector of the target house based on the business workflow, matching the address code, associating and integrating it into the house database, and using the house database to recommend house information to the user;
[0043] The search popularity of the target house and the user's interest are obtained for data integration, the target house is evaluated, the house rental and sale information is predicted based on the evaluation result, and the optimization direction of the target house is determined based on the predicted house rental and sale information.
[0044] The third aspect of the present invention also provides a computer-readable storage medium, which includes a house multi-source information management method program. When the house multi-source information management method program is executed by a processor, the steps of the house multi-source information management method as described in any one of the above items are implemented.
[0045] The present invention discloses a multi-source information management method, system and storage medium for houses, including: obtaining basic information of target houses and corresponding business requirements, extracting business workflows corresponding to business requirements, and using pre-processed basic information to fill data in business workflows; obtaining corresponding permissions according to user identity information and login information, and generating user feature vectors through user identity information and house interaction data; extracting house feature vectors of target houses based on business workflows and storing them in a house database, and recommending house information to users; obtaining the search popularity of target houses and user interest for data integration, and performing evaluation, and predicting house rental and sale information according to the evaluation results to determine the optimization direction of the target house. The present invention realizes the information management of multi-source data of houses, facilitates the release and query of house information, reduces time and management costs while managing accurately, and provides data support for the rental and sale of houses. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] Figure 1 A flow chart of a housing multi-source information management method of the present invention is shown;
[0047] Figure 2 A flow chart of a method for recommending housing information to a user using a housing database is shown in the present invention;
[0048] Figure 3 A flow chart of a method for predicting house rental and sale information based on house evaluation results according to the present invention is shown;
[0049] Figure 4 A block diagram of a housing multi-source information management system of the present invention is shown. DETAILED DESCRIPTION
[0050] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.
[0051] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.
[0052] Figure 1 A flow chart of a housing multi-source information management method of the present invention is shown.
[0053] like Figure 1 As shown, the first aspect of the present invention provides a housing multi-source information management method, comprising:
[0054] S102, obtaining basic information of the target house and corresponding business requirements, preprocessing the basic information, extracting the business workflow corresponding to the business requirements, and filling the business workflow with data using the preprocessed basic information;
[0055] S104, obtaining corresponding permissions according to the user's identity information and login information, obtaining corresponding pages and house data based on the permissions, obtaining the user's house interaction data, and generating a user feature vector through the user's identity information and house interaction data;
[0056] S106, extracting the house feature vector of the target house based on the business workflow, matching the address code, associating and integrating it into a house database, and using the house database to recommend house information to the user;
[0057] S108, obtaining the search popularity of the target house and the user's interest, performing data integration, evaluating the target house, predicting the house rental and sale information according to the evaluation result, and determining the optimization direction of the target house according to the predicted house rental and sale information.
[0058] It should be noted that the basic information of the target house, such as the location and apartment type, is extracted, and it is determined whether the target house is a newly added house based on the basic information; different information acquisition templates are generated for the target house based on the determination result; multi-source information of the target house is collected or updated based on the information acquisition template; data cleaning is performed on the collected or updated multi-source information to match the business needs such as renting or selling corresponding to the target house; corresponding functional requirements, data requirements and non-functional requirements are obtained for the target house based on the business needs, and a business workflow corresponding to the business needs is generated; the multi-source information of the target house needs to be verified using the data of each node in the business workflow, and if the verification is incorrect, When the multi-source information is collected again, the business workflow is filled with the multi-source data that has passed the verification; when the target house is not a new house, the deviation between the updated multi-source information and the original multi-source information is used to compensate the data for the workflow node, and the main process path corresponding to the original multi-source information is extracted. The business demand expansion, such as renting to selling, etc., is analyzed through the multi-source information deviation; according to the business demand expansion, the starting point is located in the main process path, and the historical process path corresponding to the business demand and the same predecessor task is retrieved through the example, that is, the historical process path is aggregated to remove redundant nodes, and the sub-process path corresponding to the business demand expansion is obtained, and the main process path is expanded and updated. According to the organizational structure set of the workflow node in the main process path, the business definition sub-process path is expanded to avoid too many branch judgments and other processing logic in a main process or sub-process. In the sub-process path, the subordinate organization can inherit the sub-process path defined by the superior organization, reducing the repeated setting of the process path.
[0059] Figure 2 The flowchart of the method for recommending house information to users by using a house database of the present invention is shown.
[0060] According to an embodiment of the present invention, the housing database is used to recommend housing information to the user, specifically:
[0061] S202, obtaining user vector features in a housing database, performing clustering processing on the user vector features, dividing the user vector features into a plurality of clusters, calculating a silhouette coefficient of each cluster, and determining an optimal number of clusters corresponding to the user vector features according to the silhouette coefficient;
[0062] S204, selecting an initial cluster center according to the number of clusters, determining clustering results of different user categories according to iterative clustering and division, and obtaining average user characteristics of different user categories according to the clustering results;
[0063] S206, obtaining a user vector feature of the current user, obtaining a category label of the current user according to the comparison between the user vector feature and the average features of users of different categories, and importing the user vector feature into the LSTM model to learn the temporal features of user preferences from the house interaction data;
[0064] S208, capturing important house interaction data in the time series features through a self-attention network, obtaining weight information of the important house interaction data, and weighting the time series features using the weight information to obtain the preference features of the current user;
[0065] S210, constructing an undirected graph by graph representation of the data in the housing database, learning the undirected graph through a graph convolutional network, and generating a graph structure according to a user vector feature of the current user;
[0066] S212, obtaining similar users through the category label of the current user, generating a neighbor matrix according to the similar users and the house feature vectors corresponding to the preference features, and using the neighbor matrix to represent the graph structure;
[0067] S214, learning the graph structure by combining the message propagation and neighbor aggregation mechanism of the graph convolutional neural network, updating the user vector feature and the house feature vector, obtaining the inner product of the updated user vector feature and the house feature vector, and recommending house information for the current user based on the inner product;
[0068] It should be noted that the silhouette coefficient of each user vector feature i is calculated in each cluster The calculation formula of its silhouette coefficient is: ,in, represents the average distance from user vector feature i to other points in the cluster, Represents the average distance between user vector feature i and all points in adjacent clusters. The silhouette coefficients of all user vector features in the cluster are averaged to obtain the total silhouette coefficient of the clustering result. The larger the silhouette coefficient, the better the clustering effect. The optimal number of clusters is determined based on the silhouette coefficient.
[0069] Obtain user identity information and login information, extract the user's login request, verification token and terminal address information based on the identity information and login information, and determine whether the user is a legitimate user based on the verification token; when the user is a legitimate user, determine whether the user is visiting for the first time, edit the user profile information and improve the identity information based on the determination result, query the user's authority level through the identity information to generate the corresponding access page and house information; obtain the house interaction data of the user on the access page, set an interaction threshold, use the interaction threshold to filter the house interaction data, and obtain the weight information of the house interaction data based on the interaction frequency and interaction duration; combine the filtered house interaction data with the identity information for feature encoding, introduce the attention mechanism in the encoder network to set the attention weight for the house interaction data, combine the weight information to perform feature fusion on the encoded features, and generate a user feature vector through the decoder;
[0070] It should be noted that the two-dimensional map information of the target area is obtained, the two-dimensional map information is gridded, each grid is encoded, the house information in the grid is counted, the house information is classified and encoded using business data, and the address coding criteria of the house are generated; the multi-source information after the target house is matched is determined according to the business workflow, the position representation data in the multi-source information after matching is extracted, the position representation data is formatted and coordinate converted, and the corresponding address code is generated as an identification code through the address coding criteria; the non-position representation data in the multi-source information is projected and transformed to generate the graph representation data of the multi-source information, the address code is combined with the generated graph representation data to generate the house feature vector of the target house, and stored in the house database. In addition, the house database provides functional interfaces including house safety, market transactions, housing security, housing rental, property management, etc. for users to use, so as to realize the scientific information management of house data and improve the level of public services.
[0071] Figure 3 The flowchart of the method for predicting house rental and sale information based on house evaluation results of the present invention is shown.
[0072] According to an embodiment of the present invention, the search popularity of the target house and the user's interest are obtained for data integration, the target house is evaluated, the house rental and sale information is predicted according to the evaluation result, and the optimization direction of the target house is determined according to the predicted house rental and sale information, specifically:
[0073] S302, obtaining the search popularity of the target house according to the number of views, recommendations, and promotions of the target house within a preset time and the time interval between the first and last views of the user, and obtaining the weight information of the target house in the user preference feature according to the user feature vector in the house database within the preset time;
[0074] S304, performing average processing on the weight information of each customer in different user categories, obtaining the interest of different user categories in the target house, and correcting the interest of different user categories in the target house by using the average transaction time of the target house;
[0075] S306, introducing attention weights to evaluate the target house by weighted calculation with the search popularity of the target house and the user's interest, building a house rental and sale prediction model based on a deep learning method, and training through house rental and sale examples;
[0076] S308, using the evaluation result of the target house as the input of the house rental and sale prediction model to obtain the predicted rental and sale period of the target house, and when the predicted rental and sale period is greater than a preset time threshold, similar instances whose rental and sale periods meet the standard threshold are screened according to the house rental and sale instance;
[0077] S310: Determine an optimization direction of the target house according to the deviation between the similar instance and the target house.
[0078] It should be noted that the time interval between the user's first browsing and the last browsing reflects the time span and continuous demand trend of the user's corresponding business needs. The interest of different user categories in the target house is corrected by the average transaction time of the target house, so as to capture the consumption habits of different categories of users, avoid the defect of large randomness of preference information, and more accurately identify the value of different categories of users. Preferably, a house rental and sale prediction model is constructed based on deep learning methods such as neural networks and support vector machines, and the target house is optimized according to the prediction results, such as adding furniture, changing the decoration style, etc.
[0079] When the target house is not rented or sold within the predicted rental and sale cycle, a similarity judgment is made based on the house feature vector of the target house and the user's preference features, and a set of users to be recommended is obtained based on the similarity. Users who have interacted with the target house are screened from the set of users to be recommended and marked, and the highest promotion priority is set based on the marked users to be recommended; a promotion plan is generated by extracting the focus of different user categories through the house feature vector of the target house, and promotion information is generated based on the promotion priority and promotion plan for push to users.
[0080] Figure 4 A block diagram of a housing multi-source information management system of the present invention is shown.
[0081] The second aspect of the present invention further provides a housing multi-source information management system 4, the system comprising: a memory 41, a processor 42, the memory comprising a housing multi-source information management method program, the housing multi-source information management method program when executed by the processor to implement the following steps:
[0082] Obtain basic information of the target house and corresponding business requirements, perform data preprocessing on the basic information, extract the business workflow corresponding to the business requirements, and use the preprocessed basic information to fill in data for the business workflow;
[0083] Obtain corresponding permissions according to the user's identity information and login information, obtain corresponding pages and house data based on the permissions, obtain the user's house interaction data, and generate a user feature vector through the user's identity information and house interaction data;
[0084] Extracting the house feature vector of the target house based on the business workflow, matching the address code, associating and integrating it into the house database, and using the house database to recommend house information to the user;
[0085] The search popularity of the target house and the user's interest are obtained for data integration, the target house is evaluated, the house rental and sale information is predicted based on the evaluation result, and the optimization direction of the target house is determined based on the predicted house rental and sale information.
[0086] The third aspect of the present invention also provides a computer-readable storage medium, which includes a house multi-source information management method program. When the house multi-source information management method program is executed by a processor, the steps of the house multi-source information management method as described in any one of the above items are implemented.
[0087] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.
[0088] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.
[0089] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.
[0090] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.
[0091] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.
[0092] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.
Claims
1. A housing multi-source information management method, characterized in that: The following steps are involved: Obtain basic information of the target house and corresponding business requirements, perform data preprocessing on the basic information, extract the business workflow corresponding to the business requirements, and use the preprocessed basic information to fill in data for the business workflow; Obtain corresponding permissions according to the user's identity information and login information, obtain corresponding pages and house data based on the permissions, obtain the user's house interaction data, and generate a user feature vector through the user's identity information and house interaction data; Extracting the house feature vector of the target house based on the business workflow, matching the address code, associating and integrating it into the house database, and using the house database to recommend house information to the user; Obtain the search popularity of the target house and the user's interest for data integration, evaluate the target house, predict the house rental and sale information according to the evaluation result, and determine the optimization direction of the target house according to the predicted house rental and sale information; The housing database is used to recommend housing information to the user, specifically: Acquire user vector features in a housing database, perform clustering on the user vector features, divide the user vector features into a plurality of clusters, calculate a silhouette coefficient of each cluster, and determine an optimal number of clusters corresponding to the user vector features according to the silhouette coefficient; Selecting an initial cluster center according to the number of clusters, determining clustering results of different user categories according to iterative clustering and division, and obtaining average user characteristics of different user categories according to the clustering results; Obtain a user vector feature of the current user, obtain a category label of the current user based on the comparison between the user vector feature and the average features of users of different categories, and import the user vector feature into the LSTM model to learn the temporal features of user preferences from the house interaction data; Capturing important house interaction data in the time series features through a self-attention network, obtaining weight information of the important house interaction data, and weighting the time series features using the weight information to obtain the preference features of the current user; Representing the data in the housing database as a graph to construct an undirected graph, learning the undirected graph through a graph convolutional network, and generating a graph structure according to a user vector feature of a current user; Obtain similar users through the category label of the current user, generate a neighbor matrix according to the similar users and the house feature vectors corresponding to the preference features, and use the neighbor matrix to represent the graph structure; Through the message propagation and neighbor aggregation mechanism of the graph convolutional neural network, the graph structure is learned, the user vector feature and the house feature vector are updated, the inner product of the updated user vector feature and the house feature vector is obtained, and the house information is recommended to the current user based on the inner product; Obtain the search popularity of the target house and the user's interest for data integration, evaluate the target house, predict the house rental and sale information based on the evaluation results, and determine the optimization direction of the target house based on the predicted house rental and sale information, specifically: The search popularity of the target house is obtained based on the number of views, recommendations, and promotions of the target house within a preset time, as well as the time interval between the first and last views of the user, and the weight information of the target house in the user preference features is obtained based on the user feature vector in the house database within the preset time; The weight information of each customer in different user categories is averaged to obtain the interest of different user categories in the target house, and the interest of different user categories in the target house is corrected by using the average transaction time of the target house; Attention weights are introduced to evaluate target houses by weighted calculation based on the search popularity of the target house and the user's interest. A house rental and sale prediction model is built based on deep learning methods and trained through house rental and sale examples. The evaluation result of the target house is used as the input of the house rental and sale prediction model to obtain the predicted rental and sale period of the target house. When the predicted rental and sale period is greater than a preset time threshold, similar instances whose rental and sale periods meet the standard threshold are screened according to the house rental and sale instances; An optimization direction of the target house is determined according to the deviation between the similar instance and the target house.
2. A housing multi-source information management method according to claim 1, characterized in that: The basic information is preprocessed to extract the business workflow corresponding to the business needs, and the preprocessed basic information is used to fill the business workflow with data, specifically: Extracting basic information of a target house, judging whether the target house is a newly added house according to the basic information, generating different information acquisition templates for the target house according to the judgment result, and collecting multi-source information of the target house or updating multi-source information according to the information acquisition template; Clean the collected or updated multi-source information, match the business requirements of the target house, obtain the corresponding functional requirements, data requirements and non-functional requirements for the target house according to the business requirements, and generate the business workflow corresponding to the business requirements; The multi-source information of the target house needs to be verified using the data of each node in the business workflow. If the verification fails, the multi-source information is collected again, and the business workflow is filled with the multi-source data that passes the verification; When the target house is not a newly added house, the deviation between the updated multi-source information and the original multi-source information is used to compensate the data for the workflow node, extract the main process path corresponding to the original multi-source information, and expand the business needs through the multi-source information deviation analysis; According to the business demand expansion, the starting point is located in the main process path, and the historical process path corresponding to the business demand and with the same predecessor task is retrieved through instances. The historical process paths are aggregated to remove redundant nodes, and the sub-process paths corresponding to the business demand expansion are obtained, and the main process path is expanded and updated.
3. A housing multi-source information management method according to claim 1, characterized in that: Based on the permissions, the corresponding page and house data are obtained, the house interaction data of the user is obtained, and the user feature vector is generated through the user's identity information and house interaction data, specifically: Obtaining user identity information and login information, extracting the user's login request, verification token and terminal address information based on the identity information and login information, and determining whether the user is a legitimate user based on the verification token; If the user is a legitimate user, determine whether it is the first time for the user to visit, edit the user information and complete the identity information according to the determination result, query the user's authority level through the identity information to generate the corresponding access page and house information; Obtaining the house interaction data of the user on the visited page, setting an interaction threshold, using the interaction threshold to filter the house interaction data, and obtaining weight information of the house interaction data according to the interaction frequency and interaction duration; The filtered house interaction data is combined with the identity information for feature encoding, and the attention mechanism is introduced in the encoder network to set the attention weight for the house interaction data. The encoded features are fused in combination with the weight information, and the user feature vector is generated through the decoder.
4. A housing multi-source information management method according to claim 1, characterized in that: Based on the business workflow, the house feature vector of the target house is extracted, and the address code is matched, associated, integrated and stored in the house database. Specifically: Acquire two-dimensional map information of the target area, grid the two-dimensional map information, encode each grid, count the house information in the grid, classify and encode the house information using business data, and generate address coding criteria for the house; Determine the multi-source information after the target house is matched according to the business workflow, extract the location representation data in the matched multi-source information, perform format conversion and coordinate conversion on the location representation data, and generate a corresponding address code as an identification code according to the address coding rule; The non-position representation data in the multi-source information is projected and transformed to generate the graph representation data of the multi-source information, the address code is combined with the generated graph representation data to generate the house feature vector of the target house, and the result is stored in the house database.
5. A housing multi-source information management system, characterized in that: The system includes: a memory and a processor, wherein the memory includes a housing multi-source information management method program, and when the housing multi-source information management method program is executed by the processor, the following steps are implemented: Obtain basic information of the target house and corresponding business requirements, perform data preprocessing on the basic information, extract the business workflow corresponding to the business requirements, and use the preprocessed basic information to fill in data for the business workflow; Obtain corresponding permissions according to the user's identity information and login information, obtain corresponding pages and house data based on the permissions, obtain the user's house interaction data, and generate a user feature vector through the user's identity information and house interaction data; Extracting the house feature vector of the target house based on the business workflow, matching the address code, associating and integrating it into the house database, and using the house database to recommend house information to the user; Obtain the search popularity of the target house and the user's interest for data integration, evaluate the target house, predict the house rental and sale information according to the evaluation result, and determine the optimization direction of the target house according to the predicted house rental and sale information; The housing database is used to recommend housing information to the user, specifically: Acquire user vector features in a housing database, perform clustering on the user vector features, divide the user vector features into a plurality of clusters, calculate a silhouette coefficient of each cluster, and determine an optimal number of clusters corresponding to the user vector features according to the silhouette coefficient; Selecting an initial cluster center according to the number of clusters, determining clustering results of different user categories according to iterative clustering and division, and obtaining average user characteristics of different user categories according to the clustering results; Obtain a user vector feature of the current user, obtain a category label of the current user based on the comparison between the user vector feature and the average features of users of different categories, and import the user vector feature into the LSTM model to learn the temporal features of user preferences from the house interaction data; Capturing important house interaction data in the time series features through a self-attention network, obtaining weight information of the important house interaction data, and weighting the time series features using the weight information to obtain the preference features of the current user; Representing the data in the housing database as a graph to construct an undirected graph, learning the undirected graph through a graph convolutional network, and generating a graph structure according to a user vector feature of a current user; Obtain similar users through the category label of the current user, generate a neighbor matrix according to the similar users and the house feature vectors corresponding to the preference features, and use the neighbor matrix to represent the graph structure; Through the message propagation and neighbor aggregation mechanism of the graph convolutional neural network, the graph structure is learned, the user vector feature and the house feature vector are updated, the inner product of the updated user vector feature and the house feature vector is obtained, and the house information is recommended to the current user based on the inner product; Obtain the search popularity of the target house and the user's interest for data integration, evaluate the target house, predict the house rental and sale information based on the evaluation results, and determine the optimization direction of the target house based on the predicted house rental and sale information, specifically: The search popularity of the target house is obtained based on the number of views, recommendations, and promotions of the target house within a preset time, as well as the time interval between the first and last views of the user, and the weight information of the target house in the user preference features is obtained based on the user feature vector in the house database within the preset time; The weight information of each customer in different user categories is averaged to obtain the interest of different user categories in the target house, and the interest of different user categories in the target house is corrected by using the average transaction time of the target house; Attention weights are introduced to evaluate target houses by weighted calculation based on the search popularity of the target house and the user's interest. A house rental and sale prediction model is built based on deep learning methods and trained through house rental and sale examples. The evaluation result of the target house is used as the input of the house rental and sale prediction model to obtain the predicted rental and sale period of the target house. When the predicted rental and sale period is greater than a preset time threshold, similar instances whose rental and sale periods meet the standard threshold are screened according to the house rental and sale instances; An optimization direction of the target house is determined according to the deviation between the similar instance and the target house.
6. A housing multi-source information management system according to claim 5, characterized in that: Based on the business workflow, the house feature vector of the target house is extracted, and the address code is matched, associated, integrated and stored in the house database. Specifically: Acquire two-dimensional map information of the target area, grid the two-dimensional map information, encode each grid, count the house information in the grid, classify and encode the house information using business data, and generate address coding criteria for the house; Determine the multi-source information after the target house is matched according to the business workflow, extract the location representation data in the matched multi-source information, perform format conversion and coordinate conversion on the location representation data, and generate a corresponding address code as an identification code according to the address coding rule; The non-position representation data in the multi-source information is projected and transformed to generate the graph representation data of the multi-source information, the address code is combined with the generated graph representation data to generate the house feature vector of the target house, and the result is stored in the house database.
7. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a house multi-source information management method program, and when the house multi-source information management method program is executed by a processor, the steps of the house multi-source information management method according to any one of claims 1 to 4 are implemented.
