A multi-dimensional data intelligent analysis and evaluation system for a building

Abstract

This invention provides a multi-dimensional data intelligent analysis and evaluation system for real estate projects, belonging to the field of data analysis technology. It includes: an information extraction module for acquiring the residential needs of target customers and performing word segmentation and mapping on these needs to obtain multiple demand information of the target customers; a real estate determination module for determining the target residential location of the target customers based on the multiple demand information, and determining multiple target real estate projects based on the target residential location; a demand analysis module for analyzing each demand information based on a pre-trained demand analysis model to determine multi-dimensional demand data; and a real estate recommendation module for acquiring the real estate information of each target real estate project and calculating a demand score for each target real estate project based on the corresponding multi-dimensional demand data, and recommending real estate projects to the target customers. This achieves intelligent and accurate matching of needs and real estate projects, improving housing selection efficiency and satisfaction.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, and in particular to a multi-dimensional intelligent data analysis and evaluation system for real estate projects. Background Technology

[0002] Multi-dimensional analysis of properties is a key decision-making mechanism to ensure the health, safety and convenience of residents throughout their life cycle. Through systematic information integration, spaces that are highly matched with individual living needs can be identified in the complex urban supply, reducing the irreversible adaptation costs caused by functional mismatch or environmental risks.

[0003] Existing technologies primarily acquire multi-dimensional data such as surrounding traffic and living environment information for each property, and then score the suitability of each property based on these characteristics. However, because they fail to comprehensively consider user needs—for example, dimensions like commuting are not included in the scoring process—they lack sufficient adaptability to users, leading to reduced accuracy in the scoring.

[0004] Therefore, this invention proposes a multi-dimensional data intelligent analysis and evaluation system for real estate projects. Summary of the Invention

[0005] This invention provides a multi-dimensional data intelligent analysis and evaluation system for real estate projects, which solves the defects of existing technologies that fail to fully consider user needs, resulting in insufficient adaptability to users and thus reduced accuracy of scoring.

[0006] This invention provides a multi-dimensional data intelligent analysis and evaluation system for real estate developments, comprising:

[0007] The information extraction module is used to obtain the housing needs of the target customers, and to perform word segmentation and word mapping on the housing needs to obtain multiple demand information of the target customers;

[0008] The property selection module is used to determine the target residence of the target customer based on the multiple demand information, and to determine multiple target properties based on the target residence;

[0009] The requirements analysis module is used to analyze each requirement information based on a pre-trained requirements analysis model and determine the corresponding multi-dimensional requirements data for each requirement information.

[0010] The property recommendation module is used to obtain property information for each target property and calculate a demand score for each target property based on the corresponding multi-dimensional demand data for each demand information, and then recommend properties to the target customers.

[0011] Preferably, the information extraction module includes:

[0012] The word segmentation processing unit is used to convert the housing demand into text to obtain demand text, and to perform word segmentation processing and word segmentation mapping on the demand text to determine the matching set of each word.

[0013] An initial construction unit is used to obtain the word attributes of each mapped matching word, take the attribute with the highest frequency of word attributes as the reference attribute, and construct the initial line based on the reference attribute, the ontology attribute of the word segmentation, and the number of mapped matching words in the matching set;

[0014] Anchor point optimization unit is used to optimize the initial route to obtain the required route based on the interest anchor points of each mapped matching word in the word segmentation and the matching set.

[0015] The forward analysis unit is used to perform forward correlation analysis between the route theme of each demand route and the route theme of the other demand routes, and to determine the connection nodes and connection weights between the demand route and each other route to obtain the route map.

[0016] The comprehensive analysis unit is used to perform comprehensive analysis on the route map to obtain multiple demand information of the target customer.

[0017] Preferably, the property determination module includes:

[0018] The demand filtering unit is used to filter sub-information related to the place of residence from each demand information;

[0019] The initial residence determination unit is used to determine multiple initial target residences for target customers based on each piece of sub-information;

[0020] The residence determination unit is used to determine the overlapping area as the target residence of the target customer when there is an overlap of multiple initial target residences;

[0021] When there are no overlapping areas, priority is determined based on the importance weight of sub-information, and the initial target residence or its extended range with the highest priority is determined as the target residence;

[0022] If it still cannot be determined, trigger a second confirmation from the customer.

[0023] Preferably, the property determination module further includes:

[0024] The initial property identification unit is used to identify multiple properties within the target residential area as the initial properties for the target residential area.

[0025] The target property determination unit is used to obtain property information for each initial property and determine whether there are any unsold units in each initial property. If so, the initial property is determined as the target property.

[0026] Preferably, the requirements analysis module includes:

[0027] The model training unit is used to train the neural network model using historical demand information and historical multi-dimensional demand data to obtain a pre-trained demand analysis model.

[0028] The model analysis unit is used to input each current requirement information into a pre-trained requirement analysis model and output multi-dimensional requirement data corresponding to each current requirement information.

[0029] Preferably, the property recommendation module includes:

[0030] The matching degree calculation unit is used to obtain the property information of each target property and calculate the first matching degree between the property information of each target property and the corresponding multi-dimensional demand data of each demand information.

[0031] The property filtering unit is used to filter target properties based on the first matching degree, resulting in multiple first-level properties;

[0032] The property identification unit is used to identify multiple second properties based on the property information of each first property and the multi-dimensional demand data, and to calculate the score of each second property for property recommendation to target customers.

[0033] Preferably, the matching degree calculation unit includes:

[0034] The information acquisition block is used to retrieve the property information of each target property from the property information database;

[0035] The information quantification block is used to quantify the property information to obtain multi-dimensional property data;

[0036] The second matching degree calculation block is used to calculate the second matching degree of each dimension between the multi-dimensional demand data corresponding to each demand information and the multi-dimensional property data.

[0037] The first matching degree calculation block is used to determine the weight of each dimension based on the demand information, and calculate the first matching degree of each target property according to the weight of each dimension and the second matching degree of each dimension.

[0038] The property recommendation block is used to identify properties whose first match score exceeds the preset match score as the first property.

[0039] Preferably, the property determination unit includes:

[0040] The search block is used to search for each requirement information based on the initial benchmark of each set dimension to obtain the first information of each sub-information in the corresponding requirement information based on each initial benchmark.

[0041] The minimum quantity determination block is used to perform cluster analysis on the multi-dimensional demand data corresponding to all demand information to obtain clusters. Based on the first number of sub-information involved in the corresponding demand information, the second number of the set dimensions involved, and the maximum number of sub-information involved under the same set dimension, the minimum number of parameters for the corresponding cluster is determined.

[0042] The parameter combination block is used to map the first information to the clusters according to the matching relationship between the corresponding clusters and the set dimensions, obtain the required parameter combination of the corresponding clusters according to the minimum number of parameters, and transform it to obtain the representative required data of the corresponding dimension.

[0043] The demand filtering block is used to filter all first-level properties to obtain third-level properties based on representative demand data for each dimension.

[0044] The frequency filtering block is used to extract third properties that appear more than a preset number of times from all third properties corresponding to the representative demand data of each dimension, and identify them as second properties.

[0045] The planning data determination block is used to obtain urban planning information and determine multi-dimensional urban planning data;

[0046] The development determination block is used to determine the planned development value and planned development weight of each second property based on multi-dimensional urban planning data.

[0047] The third matching degree calculation block is used to construct a demand matrix based on the representative demand data of each dimension, and calculate the third matching degree between each second property and the demand matrix;

[0048] The scoring calculation block is used to calculate the score of each second property based on its planned development value, planned development weight, and third-party matching degree.

[0049]

[0050] Where Score is the rating of the corresponding second property, n is the number of dimensions in the demand matrix, and ω i For the weight of the i-th dimension, SD i The second matching degree is the i-th dimension. The third matching degree is represented by PV, which is the planned development value of the corresponding second property, PW, which is the planned development weight of the corresponding second property, α, which is the first weight coefficient, and β, which is the second weight coefficient. max γ represents the maximum planned development value of all second-tier developments, where γ is the scaling factor.

[0051] The recommendation block is used to recommend second properties to target customers whose ratings exceed preset expectations.

[0052] Compared with the prior art, the beneficial effects of this application are as follows:

[0053] By accurately extracting customer housing needs, breaking down multi-dimensional indicators, matching property information, and quantifying scores, we not only meet customer needs but also achieve intelligent and precise matching between needs and properties, improving the efficiency and satisfaction of house selection.

[0054] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0055] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0056] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0057] Figure 1 This is a structural diagram of a multi-dimensional data intelligent analysis and evaluation system for real estate projects, as described in an embodiment of the present invention. Detailed Implementation

[0058] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0059] This invention provides a multi-dimensional data intelligent analysis and evaluation system for real estate projects, such as... Figure 1 As shown, it includes:

[0060] The information extraction module is used to obtain the housing needs of the target customers, and to perform word segmentation and word mapping on the housing needs to obtain multiple demand information of the target customers;

[0061] The property selection module is used to determine the target residence of the target customer based on the multiple demand information, and to determine multiple target properties based on the target residence;

[0062] The requirements analysis module is used to analyze each requirement information based on a pre-trained requirements analysis model and determine the corresponding multi-dimensional requirements data for each requirement information.

[0063] The property recommendation module is used to obtain property information for each target property and calculate a demand score for each target property based on the corresponding multi-dimensional demand data for each demand information, and then recommend properties to the target customers.

[0064] In this embodiment, the target customer refers to a potential user with a need to buy or rent a house. Housing needs are the target customer's expectations for housing and surrounding conditions, which are usually expressed in natural language, such as: want to buy a two-bedroom apartment with a balcony near the city's central hospital, and have parking in the community.

[0065] Word segmentation is the process of breaking down natural language text related to housing needs into independent words or phrases. It is a fundamental step in natural language processing. For example, the sentence "a two-bedroom apartment near the city center hospital with parking" is segmented into the following words: "near the city center hospital", "two-bedroom apartment", and "with parking". Specifically, this is achieved using NLP tools and combined with a residential domain dictionary to optimize the segmentation accuracy.

[0066] Multiple demand information are structured demands obtained after word segmentation and mapping. They are precise extractions of the original residential demands. For example, the demand information extracted from the above residential demands is: distance from the city center hospital ≤ 2 kilometers, apartment type is two bedrooms, and there are parking spaces in the community.

[0067] The target residential area is a geographical region defined based on multiple demand information, and it is the spatial range that meets the core needs.

[0068] The target property is a specific property located within the target residential area that meets the criteria of having available units for sale or rent.

[0069] The pre-trained demand analysis model is a machine learning model (usually a neural network) trained on historical demand data. It can break down single demand information into multi-dimensional detailed indicators. The model trained with 100,000 historical data (such as the distance to the hospital, hospital level, walking time, etc.) can automatically parse the dimensions of new demands. It uses the TensorFlow / PyTorch framework, takes historical demand information as input and manually labeled multi-dimensional data as output, and trains the neural network.

[0070] Each requirement is a single structured requirement obtained by the information extraction module, such as: distance from the city center hospital ≤ 2 kilometers.

[0071] Multidimensional demand data are quantitative indicators that refine individual demand information from different perspectives, reflecting the multi-layered requirements of the demand.

[0072] Property information is detailed attribute data of the target property, covering basic information, surrounding amenities, etc. Specifically, it involves crawling data from real estate databases or the property's official website, including location (latitude and longitude), unit type, amenities (medical, educational, transportation), property status, etc.

[0073] The demand score is a comprehensive score calculated based on the degree of matching between property information and multi-dimensional demand data, reflecting the extent to which the property meets the demand.

[0074] The property recommendation system presents properties to target customers in descending order of their needs based on their ratings, prioritizing properties with higher ratings.

[0075] The beneficial effects of the above technical solution are: by accurately extracting customer housing needs, breaking down multi-dimensional indicators, matching property information and quantifying scores, it not only meets customer needs, but also achieves intelligent and accurate matching between needs and properties, improving housing selection efficiency and satisfaction.

[0076] This invention provides a multi-dimensional data intelligent analysis and evaluation system for real estate developments, wherein the information extraction module includes:

[0077] The word segmentation processing unit is used to convert the housing demand into text to obtain demand text, and to perform word segmentation processing and word segmentation mapping on the demand text to determine the matching set of each word.

[0078] An initial construction unit is used to obtain the word attributes of each mapped matching word, take the attribute with the highest frequency of word attributes as the reference attribute, and construct the initial line based on the reference attribute, the ontology attribute of the word segmentation, and the number of mapped matching words in the matching set;

[0079] Anchor point optimization unit is used to optimize the initial route to obtain the required route based on the interest anchor points of each mapped matching word in the word segmentation and the matching set.

[0080] The forward analysis unit is used to perform forward correlation analysis between the route theme of each demand route and the route theme of the other demand routes, and to determine the connection nodes and connection weights between the demand route and each other route to obtain the route map.

[0081] The comprehensive analysis unit is used to perform comprehensive analysis on the route map to obtain multiple demand information of the target customer.

[0082] In this embodiment, text conversion is the process of converting non-textual housing needs (such as voice or handwritten notes) into standardized text, which is the basis for subsequent processing. The demand text is the standardized text content obtained after text conversion.

[0083] Word segmentation mapping associates each word with a pre-defined residential demand lexicon to find semantically related matching words and clarify the core meaning of the word segmentation. For example, "next to a school" is mapped to "within a distance range" or "close to a school" in the lexicon; "can park" is mapped to "has parking spaces" or "parking facilities". Specifically, by constructing a real estate lexicon (including major categories and sub-categories such as transportation, education, and apartment types), word segmentation is mapped to the lexicon through word vector similarity calculation.

[0084] The matching set of each segmented word is the set of all relevant matching words obtained after mapping each segmented word, reflecting the semantic expansion range of the segmented word. For example, the matching set of "big three-bedroom apartment" is {"three-bedroom apartment", "3-room 1-living-room", "area ≥ 120 square meters", "suitable for a family of three"}.

[0085] The word attributes of the mapped matching words are the characteristic attributes of each word in the matching set, including word type (such as noun, adjective), the category of the demand it belongs to (such as housing type, transportation), the degree of quantification (such as near, far), etc. For example, the attributes of "near the school" are {"category: educational facilities; quantification: short distance"}, and the attributes of "3-room 1-living-room" are {"category: housing type; quantification: number of rooms is 3"}.

[0086] The reference attribute is the attribute with the highest occurrence frequency selected from the word attributes of the mapped matching words, serving as the core basis for constructing the initial route. For example, the matching set of "next to the school" is {"distance to the school ≤ 1 km", "near the school"}, and both of their word attributes include educational facilities, so educational facilities is the reference attribute.

[0087] The inherent attribute of a segmented word is the attribute inherent to the segmented word itself, different from the attributes of the matching words, reflecting the most direct semantic features of the segmented word. For example, the inherent attribute of the segmented word "not too far to go to work" is {"category: commuting; core: short distance"}.

[0088] The number of mapped matching words in the matching set is the total number of matching words contained in the matching set of each segmented word. The more the number, the richer the semantics of the segmented word, which can be used to adjust the coverage range of the initial route. For example, the matching set of "can park" has 3 words ("has parking space", "parking lot supporting facilities", "sufficient parking spaces"), and the number is 3; the matching set of "big three-bedroom apartment" has 4 words, and the number is 4.

[0089] The initial route is a preliminary demand association path constructed based on the reference attribute, the inherent attribute of the segmented word, and the number of matching words, reflecting the demand logic of a single segmented word. For example, for the segmented word "next to the school", combining the reference attribute: educational facilities, the inherent attribute: near the school, and the number of matching words 3, the initial route is constructed as: educational facilities → near the school → including educational resources.

[0090] The interest anchor point is the most core appeal point in the segmented word and its matching words, which is the keyword of the customer's demand and determines the accuracy of the demand. Specifically: identify the high-frequency core words from the segmented word and the matching set through keyword extraction algorithms (such as TF-IDF, TextRank), and determine the interest anchor point in combination with domain rules.

[0091] The demand route is the accurate demand path obtained after optimizing the initial route, integrating the interest anchor point to clarify the demand boundary. For example, based on the interest anchor point "distance to the school ≤ 1 km", the initial route: educational facilities → near the school → including educational resources is optimized to: educational facilities → distance to the school ≤ 1 km → corresponding.

[0092] The theme of each demand route is the core demand category of each demand route. It is a label for the demand route and facilitates the analysis of the relationship between demands. For example, the theme of a demand route is education demand: Education facilities → distance from school ≤ 1 km → corresponding school; Commuting → commuting time ≤ 30 minutes → direct subway access. The theme of a demand route is transportation demand.

[0093] Positive correlation analysis analyzes the correlation (such as complementary, causal, or parallel relationships) between the themes of different demand routes, and identifies the inherent connections between demands. For example, the demand for education (close to school) and the demand for housing type (large three-bedroom apartment) have a complementary relationship that is suitable for families with children; the demand for transportation (direct subway access) and the demand for commuting (close to work) have a causal relationship.

[0094] Connecting nodes are the points of connection between different demand routes, that is, they meet the common conditions of multiple routes at the same time. For example, the connecting node between the education demand route (≤1 km from the school) and the living support demand route (supermarket nearby) is an area within 1 km of the school and within 500 meters of a supermarket.

[0095] The connection weight is a numerical value (range 0-1) that measures the strength of the connection between lines with different needs. The higher the weight, the stronger the connection. The specific calculation method is as follows:

[0096] Where Wrj represents the degree of association between demand route r and demand route j (r≠j); Sim(Tr,Tj) represents the semantic similarity of the route topics of demand route r and demand route j; Po(Lr,Lj) represents the path overlap between demand route r and demand route j, and

[0097] NM() is the node matching function; Nrk represents the k-th node of demand route r, and Njk represents the k-th node of demand route j; μ1 and μ2 are weight coefficients, and μ1+μ2=1 is used to balance the influence of semantic similarity and path overlap. If the demand is mainly based on vague topics, μ1 is increased; if the demand is mainly based on specific nodes, μ2 is increased; OD(Lr,Lj) is the abnormal correlation degree between demand route r and demand route j. IO() is the function for judging abnormal nodes; GO is the global anomaly baseline value; μ3 is the correction coefficient, and μ3>0; UP(Lr,Lj) is the user preference coefficient, and... Click(Lr,Lj) represents the number of clicks based on demand route r and demand route j; Collect(Lr,Lj) represents the number of collections based on demand route r and demand route j; τ1 is the collection weight coefficient; TI is the total number of interactions; MUP is the global maximum user preference value; δ0 is the preference adjustment coefficient, and δ0≥0, amplifying the influence of user preference on the weight.

[0098] A route map is a visual association graph that integrates all demand routes, connecting nodes, and connecting weights. It intuitively shows the overall logic between demands. For example, a route map contains three routes: education demand, transportation demand, and housing demand. These routes are connected by connecting nodes within 5 kilometers of the city center, and the connecting weights of each node are marked.

[0099] The multiple needs of target customers are structured and quantifiable core needs extracted after comprehensive analysis of the roadmap. They serve as the direct basis for subsequent property matching. Based on the connecting nodes and weights of the roadmap, the K-means clustering algorithm is used to merge duplicate needs and retain high-weight core needs to form structured information.

[0100] The beneficial effects of the above technical solution are as follows: through word segmentation, mapping matching, route construction and optimization, association analysis and comprehensive extraction, this module transforms the customer's vague original housing needs into multiple structured and quantifiable demand information, which greatly improves the accuracy of demand analysis, provides a reliable basis for subsequent property matching and recommendation, and reduces recommendation errors caused by misunderstanding of demand.

[0101] This invention provides a multi-dimensional data intelligent analysis and evaluation system for real estate projects, wherein the project identification module includes:

[0102] The demand filtering unit is used to filter sub-information related to the place of residence from each demand information;

[0103] The initial residence determination unit is used to determine multiple initial target residences for target customers based on each piece of sub-information;

[0104] The residence determination unit is used to determine the overlapping area as the target residence of the target customer when there is an overlap of multiple initial target residences;

[0105] When there are no overlapping areas, priority is determined based on the importance weight of sub-information, and the initial target residence or its extended range with the highest priority is determined as the target residence;

[0106] If it still cannot be determined, trigger a second confirmation from the customer.

[0107] An initial target residence is a geographical area defined based on a single piece of sub-information. It is a preliminary area that satisfies that sub-information. Multiple initial target residences are multiple geographical areas output by the initial residence determination unit, such as two areas: within 5 kilometers of the workplace and within 800 meters of the market. Overlapping areas are the intersection of multiple initial target residences, that is, areas that simultaneously satisfy multiple pieces of sub-information. For example, if the areas within 5 kilometers of the workplace and within 800 meters of the market overlap, the overlapping part is the area that simultaneously satisfies the conditions of being close to both the workplace and the market.

[0108] The importance weight of sub-information is a numerical value (range 0-1) that measures the importance of each sub-information in the customer's needs. The higher the weight, the higher the priority of the initial place of residence corresponding to the sub-information. It is generally set manually by the customer. The priority is based on the ranking of the initial target place of residence according to the importance weight. The initial place of residence corresponding to the sub-information with the highest weight is considered first.

[0109] The expanded area is the area that is expanded by a preset ratio (e.g., 1.2 times the original area) when there are no suitable properties available in the highest priority initial residence location.

[0110] The beneficial effects of the above technical solution are: by filtering sub-information related to the place of residence and determining the target place of residence in different scenarios (overlapping areas are given priority, weight ranking is secondary, and customer confirmation is a last resort), it not only ensures the matching degree between the target place of residence and the core needs of customers, but also solves the problem of non-overlapping areas through flexible expansion and interaction mechanisms, thereby improving the stability and practicality of the system in complex demand scenarios.

[0111] This invention provides a multi-dimensional data intelligent analysis and evaluation system for real estate projects, wherein the project identification module further includes:

[0112] The initial property identification unit is used to identify multiple properties within the target residential area as the initial properties for the target residential area.

[0113] The target property determination unit is used to obtain property information for each initial property and determine whether there are any unsold units in each initial property. If so, the initial property is determined as the target property.

[0114] In this embodiment, the initial properties refer to all properties located within the target residential area that have been selected by the initial property determination unit, and these properties form the basis for subsequent screening.

[0115] In this embodiment, "houses for sale" refers to properties in a development that are available for sale, including new homes and secondhand homes where the landlord has explicitly stated that they are for sale.

[0116] The beneficial effects of the above technical solution are: first, multiple buildings in the target residential area are identified as the initial buildings in the target residential area, then the building information of each initial building is obtained, and the initial buildings with unsold houses are identified as the target buildings, which can ensure that there are unsold houses in the target buildings.

[0117] This invention provides a multi-dimensional data intelligent analysis and evaluation system for real estate projects. The demand analysis module includes:

[0118] The model training unit is used to train the neural network model using historical demand information and historical multi-dimensional demand data to obtain a pre-trained demand analysis model.

[0119] The model analysis unit is used to input each current requirement information into a pre-trained requirement analysis model and output multi-dimensional requirement data corresponding to each current requirement information.

[0120] The beneficial effect of the above technical solution is that it can better capture the multi-dimensional demand data corresponding to the demand information based on the model.

[0121] This invention provides a multi-dimensional data intelligent analysis and evaluation system for real estate projects, wherein the real estate recommendation module includes:

[0122] The matching degree calculation unit is used to obtain the property information of each target property and calculate the first matching degree between the property information of each target property and the corresponding multi-dimensional demand data of each demand information.

[0123] The property filtering unit is used to filter target properties based on the first matching degree, resulting in multiple first-level properties;

[0124] The property identification unit is used to identify multiple second properties based on the property information of each first property and the multi-dimensional demand data, and to calculate the score of each second property for property recommendation to target customers.

[0125] In this embodiment, the first matching degree is the overall degree of fit between the target property information and the multi-dimensional demand data, represented by a score of 0-100 or 0-1. The higher the score, the higher the matching degree. The first matching degree = ∑(dimensional weight × dimensional matching score).

[0126] The beneficial effects of the above technical solution are as follows: by calculating the matching degree, the first property that initially meets the needs is selected, and then the second property is determined and recommended after multi-dimensional needs verification. The objectivity of the selection is ensured by quantifying the matching degree, and the accuracy of the recommendation is improved by multiple rounds of selection. In the end, the property options that meet the core needs of the customers are provided, thereby improving the efficiency and satisfaction of house selection.

[0127] This invention provides a multi-dimensional data intelligent analysis and evaluation system for real estate projects, including a matching degree calculation unit, comprising:

[0128] The information acquisition block is used to retrieve the property information of each target property from the property information database;

[0129] The information quantification block is used to quantify the property information to obtain multi-dimensional property data;

[0130] The second matching degree calculation block is used to calculate the second matching degree of each dimension between the multi-dimensional demand data corresponding to each demand information and the multi-dimensional property data.

[0131] The first matching degree calculation block is used to determine the weight of each dimension based on the demand information, and calculate the first matching degree of each target property according to the weight of each dimension and the second matching degree of each dimension; the property recommendation block is used to regard target properties whose first matching degree exceeds the preset matching degree as the first property.

[0132] In this embodiment, the property information database is a structured database that stores detailed information about various properties, including fields such as basic property attributes, surrounding amenities, and property status. It is the core source for the system to obtain property data.

[0133] In this embodiment, quantification is the process of converting unstructured or descriptive property information into calculable values ​​to facilitate subsequent matching analysis. For example, 0.8 kilometers from the subway station is quantified as 800 meters; the presence of two primary schools in the vicinity is quantified as 2; and a greening rate of 35% is directly retained as the value 35%.

[0134] In this embodiment, the preset matching degree is the minimum matching threshold set by the system, which is used to filter properties that meet basic requirements (e.g., 80 points). It is set by the system administrator according to the business scenario (e.g., 70 points for first-time homebuyers and 90 points for high-end homebuyers), and can be flexibly adjusted in the background.

[0135] The beneficial effects of the above technical solution are as follows: by obtaining and quantifying property information from the database, calculating the matching degree in different dimensions, and integrating it into an overall matching score, the first property with the highest matching degree is finally selected. This not only achieves accurate comparison between demand and property information, but also reflects the priority of customer demand through weight settings, which greatly improves the scientificity and efficiency of property selection and lays the foundation for subsequent accurate recommendations.

[0136] This invention provides a multi-dimensional data intelligent analysis and evaluation system for real estate projects, wherein the real estate project determination unit includes:

[0137] The search block is used to search for each requirement information based on the initial benchmark of each set dimension to obtain the first information of each sub-information in the corresponding requirement information based on each initial benchmark.

[0138] The minimum quantity determination block is used to perform cluster analysis on the multi-dimensional demand data corresponding to all demand information to obtain clusters. Based on the first number of sub-information involved in the corresponding demand information, the second number of the set dimensions involved, and the maximum number of sub-information involved under the same set dimension, the minimum number of parameters for the corresponding cluster is determined.

[0139] The parameter combination block is used to map the first information to the clusters according to the matching relationship between the corresponding clusters and the set dimensions, obtain the required parameter combination of the corresponding clusters according to the minimum number of parameters, and transform it to obtain the representative required data of the corresponding dimension.

[0140] The demand filtering block is used to filter all first-level properties to obtain third-level properties based on representative demand data for each dimension.

[0141] The frequency filtering block is used to extract third properties that appear more than a preset number of times from all third properties corresponding to the representative demand data of each dimension, and identify them as second properties.

[0142] The planning data determination block is used to obtain urban planning information and determine multi-dimensional urban planning data;

[0143] The development determination block is used to determine the planned development value and planned development weight of each second property based on multi-dimensional urban planning data.

[0144] The third matching degree calculation block is used to construct a demand matrix based on the representative demand data of each dimension, and calculate the third matching degree between each second property and the demand matrix;

[0145] The scoring calculation block is used to calculate the score of each second property based on its planned development value, planned development weight, and third-party matching degree.

[0146]

[0147] Where Score is the rating of the corresponding second property, n is the number of dimensions in the demand matrix, and ω i For the weight of the i-th dimension, SD i The second matching degree is the i-th dimension. The third matching degree is represented by PV, which is the planned development value of the corresponding second property, PW, which is the planned development weight of the corresponding second property, α, which is the first weight coefficient, and β, which is the second weight coefficient. max The maximum value of the planned development value of all second properties, where γ is the scaling factor; the recommendation block is used to recommend second properties with scores exceeding the preset expectations to target customers.

[0148] In this embodiment, the defined dimensions are predefined perspectives for demand analysis, such as transportation, education, commerce, and apartment type. For example, customers looking to buy a house may focus on transportation (subway / bus), education (number of schools), and apartment type (3 bedrooms / north-south facing). These three categories are the defined dimensions. Specifically, the system administrator or demand analyst manually enters or imports the dimension list through a template in the system configuration interface based on common demands in the real estate industry (referencing historical data and market research). For example, a dimension configuration table is maintained, which includes dimension name, description, and associated fields.

[0149] The initial baseline is the basic condition used for searching at the beginning of each defined dimension. For example, the initial baseline for the transportation dimension is set to be ≤1 km from the subway entrance; the initial baseline for the education dimension is set to be at least 2 public primary schools within 3 km of the community. Specifically, the basic conditions associated with each dimension are obtained through the dimension-baseline configuration lookup table.

[0150] Sub-information is the breakdown of demand information into small segments, corresponding to specific requirements in various defined dimensions. For example, sub-information can be broken down into the transportation dimension: ≤1 km from the subway station (10-minute walk), sub-information can be broken down into the education dimension: there is a public primary school within 3 km, and sub-information can be broken down into the housing type dimension: three-bedroom apartment.

[0151] The first piece of information is the basic data in the property database that meets the conditions after searching based on the initial benchmark. For example, using the benchmark of ≤1 km from the subway station, the property database is used to retrieve property A (0.8 km from the subway) and property B (0.5 km from the subway). The subway distance data of these properties is the first piece of information in the transportation dimension. Similarly, property A (with 2 public primary schools nearby) and property C (with 3 public primary schools nearby) are retrieved as the first piece of information in the education dimension. The acquisition of the first piece of information is based on the property database, which contains pre-stored dimension conditions and matching property information.

[0152] Cluster analysis groups similar, multi-dimensional demand data into a group (cluster) to find common patterns in demand, facilitating batch processing. For example, if you cluster the demand data of 100 customers and find that 20 customers are interested in proximity to subway, primary school, and 3-bedroom apartments, these 20 customers are grouped into one cluster.

[0153] Clusters are sets of requirements obtained after cluster analysis.

[0154] The first quantity is the number of sub-information items involved in a certain demand information. For example, demand information: subway station ≤ 1 km, 3-bedroom apartment, nearby primary school, the sub-information items are subway, apartment type, and education, so the first quantity is 3.

[0155] The second quantity is the number of dimensions involved in the demand information, which shows the breadth of the dimensions covered by the demand. For example, if the demand involves three dimensions: transportation, education, and housing type, the second quantity is 3.

[0156] The maximum number of sub-information under the same defined dimension is the number of the most numerous requirements for sub-information within that dimension. This is used to measure the complexity of the requirements for that dimension. For example, in the transportation dimension, some requirements only mention proximity to the subway (1 sub-information), while others mention proximity to the subway, availability of bus stops, and uncongested roads (3 sub-information). Therefore, the maximum number of requirements under the transportation dimension is 3.

[0157] The minimum number of parameters is calculated based on the first, second, and maximum number of parameters, determining the minimum number of parameters (sub-information) required for each cluster. For example, cluster A has a first parameter of 3, a second parameter of 2, and a maximum parameter of 2. This is the floor symbol.

[0158] Additional mapping involves further linking and supplementing the primary information and the requirements of the cluster, making the requirements and property data more closely aligned. For example, if the cluster requirements are: distance from the subway ≤ 1 km and ≥ 2 subway lines, and the primary information for property A is: distance from the subway 0.8 km, the additional mapping would supplement it with: number of subway lines 3, making the transportation data for property A more complete.

[0159] Demand parameter combination is a set of conditions that integrates multi-dimensional demand data according to the minimum number of parameters. For example, the minimum number of parameters for a cluster is 3, which can be combined into: Transportation: ≤1 km from the subway, ≥2 subway lines; Education: ≥2 primary schools. This is demand parameter combination.

[0160] The conversion process involves transforming the combination of demand parameters into a standardized data format that the system can recognize and use to filter properties. For example, demand parameter combinations such as "distance from the subway ≤ 1 km" and "number of subway lines ≥ 2" can be converted into SQL conditions "subway distance <= 1 AND number of subway lines >= 2".

[0161] Representative demand data refers to the core data that best represents customer needs for each dimension after conversion. For example, representative demand data for the transportation dimension are: distance from the subway ≤ 1 km, subway lines ≥ 2; for the education dimension, it is: number of primary schools ≥ 2, teaching level ≥ 2.

[0162] The representative demand data consists of the core demands of each dimension output by the parameter combination block. For example, if the first property has three properties, A, B, and C, the representative demand data can be used to filter them: distance from the subway ≤ 1 km, subway lines ≥ 2. This leaves A (0.8 km from the subway, 3 lines) and B (0.5 km from the subway, 2 lines). A and B are the third property.

[0163] The second property is selected from the third property, based on the properties that appear most frequently in the demand data across multiple dimensions.

[0164] Urban planning information refers to regional development plans published by the government (such as plans to build shopping malls, schools, or new subway lines in a certain area).

[0165] Multidimensional urban planning data is structured data that breaks down urban planning information into dimensions such as transportation, education, and commerce, making it easier to correlate and analyze with real estate data.

[0166] The planned development value is a score that assesses the future development potential of a property based on urban planning. The planned development value = ∑(planning type weight × distance coefficient × scale coefficient). The distance coefficient is higher the closer the property is to the planned location (e.g., coefficient 1 for distance ≤ 1 km, coefficient 0.6 for distance 1-3 km). The scale coefficient is higher the higher the planning level. The planning level is related to the school's educational level and transportation convenience.

[0167] The planning development weight is assigned based on the degree of impact of the plan on residential value. For example, if urban planning has a significant impact on residential life, the planning development weight is set to 0.3; if customers are more concerned about existing amenities, the weight is set to 0.2. The weight values ​​are configured in the system backend based on expert research and market data verification (such as analyzing the correlation between the increase in property prices after the plan is implemented and the plan).

[0168] A demand matrix is ​​a matrix (table) that organizes demand data from multiple dimensions. Each row represents a dimension, and each column represents a demand condition, facilitating matching and calculations with property data. For example, Table 1:

[0169] Table 1 Demand Matrix List

[0170] Dimension Requirements Weight transportation ≤1 km from the subway 0.4 transportation ≥2 subway lines 0.3 educate Number of primary schools ≥ 2 0.2 Apartment type Three-bedroom apartment 0.1

[0171] The third matching score is calculated by comparing the demand matrix with the property data. For example, property A meets the following criteria in terms of transportation: 0.8 kilometers from the subway (match score 1 point) and 3 subway lines (match score 1 point); in terms of education: 2 primary schools (match score 1 point); and the apartment type is a three-bedroom unit (match score 1 point). The formula is:

[0172] In this embodiment, the weight of the first weight coefficient, the existing demand matching degree (third matching degree), in the total score reflects the importance of the current demand matching degree. For example, it is set to 0.6 to emphasize the existing supporting facilities.

[0173] The second weighting coefficient is the weight of the planned development portion of PV×PW, reflecting the importance of future potential. For example, it is set to 0.4, which emphasizes the appreciation of the property. The sum of the first weighting coefficient and the second weighting coefficient is 1.

[0174] The scaling factor adjusts the degree of influence of the planned development value (PV) on the index. For example, setting it to 0.5 makes the score growth of properties with high PV more smooth. The value of PV ranges from 0 to 1.

[0175] PV max It is used for normalization to make PV comparable across different properties.

[0176] Specifically, retrieve α, β, and γ from the database (which can be configured and adjusted in the backend), iterate through each second property, and query PV, PW, and ω. i SD i Take the data, substitute it into the formula to calculate the score, and save the result to the property rating table.

[0177] The preset expectation value is the minimum score standard set. For example, the preset expectation value is set to 8 points.

[0178] The recommendation logic involves the system checking the property rating table, filtering out the second property with a score greater than the preset expected value, sorting them by rating from high to low, and pushing them to target customers. These customers can be reached through apps, websites, SMS, and other means.

[0179] By breaking down customer needs into multi-dimensional and granular conditions, clustering is used to find commonalities in these needs. Furthermore, by combining current amenities and future development potential, properties are accurately scored and screened. Through multiple rounds of clustering and mapping, the data on needs and properties are deeply aligned, improving recommendation efficiency and customer satisfaction.

[0180] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A multi-dimensional data intelligent analysis and evaluation system for a building, characterized in that, include: The information extraction module is used to obtain the housing needs of the target customers, and to perform word segmentation and word mapping on the housing needs to obtain multiple demand information of the target customers; The property selection module is used to determine the target residence of the target customer based on the multiple demand information, and to determine multiple target properties based on the target residence; The requirements analysis module is used to analyze each requirement information based on a pre-trained requirements analysis model and determine the corresponding multi-dimensional requirements data for each requirement information. The property recommendation module is used to obtain property information for each target property and calculate the demand score for each target property based on the corresponding multi-dimensional demand data for each demand information, and recommend properties to the target customers. The property recommendation module includes: The matching degree calculation unit is used to obtain the property information of each target property and calculate the first matching degree between the property information of each target property and the corresponding multi-dimensional demand data of each demand information. The property filtering unit is used to filter target properties based on the first matching degree, resulting in multiple first-level properties; The property identification unit is used to identify multiple second properties based on the property information of each first property and the multi-dimensional demand data, and to calculate the score of each second property for property recommendation to target customers; The matching degree calculation unit includes: The information acquisition block is used to retrieve the property information of each target property from the property information database; The information quantification block is used to quantify the property information to obtain multi-dimensional property data; The second matching degree calculation block is used to calculate the second matching degree of each dimension between the multi-dimensional demand data corresponding to each demand information and the multi-dimensional property data. The first matching degree calculation block is used to determine the weight of each dimension based on the demand information, and calculate the first matching degree of each target property according to the weight of each dimension and the second matching degree of each dimension. The property recommendation block is used to consider properties whose first match score exceeds the preset match score as the first property; The property identification unit includes: The search block is used to search for each requirement information based on the initial benchmark of each set dimension to obtain the first information of each sub-information in the corresponding requirement information based on each initial benchmark. The minimum quantity determination block is used to perform cluster analysis on the multi-dimensional demand data corresponding to all demand information to obtain clusters. Based on the first number of sub-information involved in the corresponding demand information, the second number of the set dimensions involved, and the maximum number of sub-information involved under the same set dimension, the minimum number of parameters for the corresponding cluster is determined. The parameter combination block is used to map the first information to the clusters according to the matching relationship between the corresponding clusters and the set dimensions, obtain the required parameter combination of the corresponding clusters according to the minimum number of parameters, and transform it to obtain the representative required data of the corresponding dimension. The demand filtering block is used to filter all first-level properties to obtain third-level properties based on representative demand data for each dimension. The frequency filtering block is used to extract third properties that appear more than a preset number of times from all third properties corresponding to the representative demand data of each dimension, and identify them as second properties. The planning data determination block is used to obtain urban planning information and determine multi-dimensional urban planning data; The development determination block is used to determine the planned development value and planned development weight of each second property based on multi-dimensional urban planning data. The third matching degree calculation block is used to construct a demand matrix based on the representative demand data of each dimension, and calculate the third matching degree between each second property and the demand matrix; The scoring calculation block is used to calculate the score of each second property based on its planned development value, planned development weight, and third-party matching degree. in, To correspond with the rating of the second property, The number of dimensions in the demand matrix. For the first Dimension weights For the first The second degree of matching in the dimension, The third degree of matching, To correspond with the planned development value of the second property, In order to correspond to the planning and development weight of the second property, As the first weighting coefficient, This is the second weighting coefficient. This represents the maximum planned development value of all second-tier developments. This is the scaling factor; The recommendation block is used to recommend second properties to target customers whose ratings exceed preset expectations.

2. The multi-dimensional data intelligent analysis and evaluation system for real estate projects according to claim 1, characterized in that, The information extraction module includes: The word segmentation processing unit is used to convert the housing demand into text to obtain demand text, and to perform word segmentation processing and word segmentation mapping on the demand text to determine the matching set of each word. An initial construction unit is used to obtain the word attributes of each mapped matching word, take the attribute with the highest frequency of word attributes as the reference attribute, and construct the initial line based on the reference attribute, the ontology attribute of the word segmentation, and the number of mapped matching words in the matching set; Anchor point optimization unit is used to optimize the initial route to obtain the required route based on the interest anchor points of each mapped matching word in the word segmentation and the matching set. The forward analysis unit is used to perform forward correlation analysis between the route theme of each demand route and the route theme of the other demand routes, and to determine the connection nodes and connection weights between the demand route and each other route to obtain the route map. The comprehensive analysis unit is used to perform comprehensive analysis on the route map to obtain multiple demand information of the target customer.

3. The multi-dimensional data intelligent analysis and evaluation system for real estate projects according to claim 1, characterized in that, The property identification module includes: The demand filtering unit is used to filter sub-information related to the place of residence from each demand information; The initial residence determination unit is used to determine multiple initial target residences for target customers based on each piece of sub-information; The residence determination unit is used to determine the overlapping area as the target residence of the target customer when there is an overlap of multiple initial target residences; When there are no overlapping areas, priority is determined based on the importance weight of sub-information, and the initial target residence or its extended range with the highest priority is determined as the target residence; If it still cannot be determined, trigger a second confirmation from the customer.

4. The multi-dimensional data intelligent analysis and evaluation system for real estate projects according to claim 3, characterized in that, The property identification module also includes: The initial property identification unit is used to identify multiple properties within the target residential area as the initial properties for the target residential area. The target property determination unit is used to obtain property information for each initial property and determine whether there are any unsold units in each initial property. If so, the initial property is determined as the target property.

5. The multi-dimensional data intelligent analysis and evaluation system for real estate projects according to claim 1, characterized in that, The requirements analysis module includes: The model training unit is used to train the neural network model using historical demand information and historical multi-dimensional demand data to obtain a pre-trained demand analysis model. The model analysis unit is used to input each current requirement information into a pre-trained requirement analysis model and output multi-dimensional requirement data corresponding to each current requirement information.

Images