Real estate price prediction method, device, equipment, medium and program product

Abstract

The embodiment of the invention provides a real estate price prediction method and device, equipment, a medium and a program product, and relates to the field of big data. The method comprises the following steps: acquiring attribute information of a target real estate and associated information of the target real estate, wherein the associated information comprises spatial geographic information; the spatial geographic information is used for representing spatial information of the target real estate in a preset range; generating a feature matrix according to the attribute information and the associated information; determining a target valuation model from a plurality of pre-trained valuation models according to the type of the target real estate; and inputting the feature matrix into a target valuation model to obtain a predicted price of the target real estate. The real estate price is automatically predicted by fusing the attribute information and the associated information of the real estate and using the deep learning model, so that the prediction time consumption is greatly reduced, and the prediction efficiency is improved; and the method effectively improves the prediction accuracy by integrating more comprehensive information.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of big data, and in particular to a real estate price prediction method, device, equipment, medium and program product. BACKGROUND

[0002] Valuation of real estate (such as real estate, land, etc.) plays an important role in the fields of finance, real estate transactions, etc. Accurate valuation can help banks, investors and other related parties make reasonable decisions. In particular, in the bank loan business, the valuation of real estate as collateral directly affects the loan amount, interest rate setting and risk control.

[0003] The existing real estate valuation mainly relies on manual evaluation, and professional evaluators inspect the real estate on site and combine subjective factors such as market experience and similar cases to evaluate. This method generally takes a long time, usually 3-7 days, and the comprehensive cost of manual service and time investment is high. In addition, the price of real estate is easily affected by the market and is dynamically fluctuating, and the long evaluation period is easy to cause the evaluation result to lag behind the market change, and may also cause the evaluation result to miss the time limit, ultimately leading to inaccurate evaluation result. SUMMARY

[0004] The present application provides a real estate price prediction method, device, equipment, medium and program product, which realizes automatic real estate price prediction by fusing multi-source information and utilizing the powerful learning ability of the model. This method greatly compresses the cycle of traditional manual evaluation, improves the prediction efficiency, and at the same time improves the accuracy of the prediction.

[0005] In a first aspect, the present application provides a real estate price prediction method, comprising:

[0006] Obtaining attribute information of a target real estate and associated information of the target real estate, the associated information comprising spatial geographic information; the spatial geographic information is used to represent the spatial information of the target real estate within a preset range;

[0007] Generating a feature matrix according to the attribute information and the associated information;

[0008] According to the type of the target real estate, determining a target valuation model from a plurality of pre-trained valuation models;

[0009] Inputting the feature matrix into the target valuation model to obtain a predicted price of the target real estate.

[0010] In a second aspect, the present application provides a loan amount determination method, comprising:

[0011] Determining a loan amount based on the predicted price of the real estate mortgaged by the loan user; wherein the predicted price is obtained according to the method provided in the first aspect above.

[0012] In a third aspect, the present application provides a real estate price prediction device, comprising:

[0013] an information acquisition module, configured to acquire attribute information of a target real estate and associated information of the target real estate, the associated information comprising spatial geographic information, and the spatial geographic information being used to represent spatial information of the target real estate within a preset range;

[0014] a feature matrix generation module, configured to generate a feature matrix according to the attribute information and the associated information;

[0015] a model determination module, configured to determine a target evaluation model from a plurality of pre-trained evaluation models according to a type of the target real estate;

[0016] a price prediction module, configured to input the feature matrix into the target evaluation model to obtain a predicted price of the target real estate.

[0017] In a fourth aspect, the present application provides an electronic device, comprising a processor and a memory connected to the processor in communication;

[0018] the memory stores computer execution instructions;

[0019] the processor executes the computer execution instructions stored in the memory to implement the method provided in the first aspect or the second aspect.

[0020] In a fifth aspect, the present application provides a computer readable storage medium, the computer readable storage medium storing computer execution instructions, the computer execution instructions being executed by a processor to implement the method provided in the first aspect or the second aspect.

[0021] In a sixth aspect, the present application provides a computer program product, comprising a computer program, the computer program being executed by a processor to implement the method provided in the first aspect or the second aspect.

[0022] The real estate price prediction method, device, equipment, medium and program product provided by the present application acquire attribute information and associated information of a real estate itself to evaluate the real estate from a more comprehensive perspective; then, a feature matrix representing the real estate is generated according to the acquired information, and a model corresponding to the real estate is determined from a plurality of pre-trained models, so that the model can fully utilize the ability of the model to capture specific data patterns and relationships, thereby improving the accuracy of the prediction. Finally, the feature matrix is input into the model, and the model automatically outputs the predicted price of the real estate through the powerful learning and reasoning ability of the model. The model outputs the prediction result, which greatly reduces the time-consuming of obtaining the predicted price by relying on manual evaluation, improves the prediction efficiency, and the final prediction result is more accurate through the evaluation of the real estate from a more comprehensive perspective. Attached Figure Description

[0023] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0024] Figure 1 A flowchart illustrating a method for predicting the price of real estate provided in an embodiment of this application;

[0025] Figure 2 A schematic diagram of map data within a preset range provided in this application;

[0026] Figure 3 A flowchart illustrating another method for predicting the price of real estate provided in this application embodiment;

[0027] Figure 4 A schematic diagram of the structure of a real estate price prediction device provided in an embodiment of this application;

[0028] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0029] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0030] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0031] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of the relevant data all comply with the relevant laws, regulations, and standards of the relevant countries and regions, have taken necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation access points for users to choose to authorize or refuse.

[0032] Furthermore, the technical solution involved in this application, which involves big data analysis of user information (including but not limited to personal biometrics, identity data, consumption data, asset data, electronic terminal operation data, etc.) and the use of artificial intelligence technology for automated decision-making, and makes decisions that have a significant impact on personal rights based on the results of automated decision-making, provides users with corresponding operation entry points for users to choose to agree to or reject the results of automated decision-making; if the user chooses to reject, the process will proceed to the expert decision-making process.

[0033] It should be noted that the real estate price forecasting methods, devices, equipment, media, and program products provided in this application can be used in the field of big data, or in any field other than big data. The application fields of the real estate price forecasting methods, devices, equipment, media, and program products in this application are not limited.

[0034] This application applies to scenarios involving the valuation of real estate, including but not limited to houses and land. Real estate valuation plays a crucial role in various practical applications. A typical application is in bank lending and risk management. Specifically, when applying for a loan using real estate as collateral, the bank needs to assess the value of the real estate to determine a reasonable loan amount. Furthermore, after the loan is disbursed, the bank will periodically assess the value of the real estate to monitor whether the value of the collateral has changed and to avoid post-loan risks. Another typical application is in real estate transactions. Specifically, in real estate transactions, valuing the real estate helps to assess the reasonableness of the transaction price.

[0035] Therefore, there is an urgent need to provide an efficient method for real estate price forecasting to conduct efficient and accurate real estate valuation.

[0036] Currently, real estate price forecasting typically involves appraisers conducting on-site inspections of properties, collecting information, and relying on their subjective experience, market experience, and comparable cases to analyze the collected information and arrive at a predicted price. This manual information collection and analysis is time-consuming, generally requiring 3-7 days, resulting in high labor and time costs and low forecasting efficiency. Furthermore, because real estate prices are greatly affected by environmental factors, a lengthy appraisal period may cause the appraisal results to be outdated and inaccurate. In addition, manual appraisals are prone to introducing human bias, affecting the objectivity of the forecast.

[0037] The real estate price prediction method provided in this application aims to solve the aforementioned technical problems. By acquiring the attribute information and related information of the real estate itself, and using a computer to comprehensively analyze this information, the method automatically predicts the real estate price. First, based on the attribute and related information, a feature matrix representing the real estate is generated. This feature matrix is ​​then input into a pre-trained model corresponding to the real estate. The model automatically learns and infers from the feature matrix, ultimately outputting the predicted price of the real estate. This method avoids manual evaluation of real estate, significantly reducing the time spent predicting real estate prices and improving prediction efficiency. Furthermore, by integrating multi-dimensional information, it provides a more comprehensive perspective, thereby improving the accuracy of the prediction results.

[0038] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0039] Figure 1 This is a flowchart illustrating a method for predicting real estate prices according to an embodiment of this application. The method for predicting real estate prices provided in this embodiment can be executed by an electronic device with corresponding processing capabilities. Figure 1 As shown, the method provided in this embodiment includes the following steps:

[0040] Step S101: Obtain the attribute information and related information of the target real estate.

[0041] The target real estate is real estate for which price forecasting is required, such as houses or land used as collateral.

[0042] The attribute information of the target real estate is key information describing the core characteristics, legal status, physical characteristics and management requirements of the target real estate, including but not limited to the area, location, structure, age of the building, ownership, term of rights and use of the target real estate.

[0043] The method for obtaining the attribute information of the target real estate can be as follows: the electronic device executing this method establishes a communication connection with the server of the relevant department, sends a request to the server of the relevant department to obtain the attribute information of the target real estate, and the server of the relevant department returns the attribute information of the target real estate to the electronic device. It should be noted that this method of acquisition is implemented under the premise of full authorization from the owner of the target real estate and the relevant department.

[0044] Another method for obtaining the attribute information of the target real estate is to send an attribute information retrieval notification to the owner of the target real estate. Upon receiving the notification, the owner uploads the attribute information of the target real estate, along with relevant supporting documents. After receiving the uploaded attribute information and relevant supporting documents, the electronic device verifies the attribute information to determine its authenticity.

[0045] Related information about the target property refers to external information that is directly or indirectly related to the target property and affects its rights exercise, value assessment, transaction security, or management and supervision. Although related information is not part of the target property's own attribute information, it has a significant impact on the target property's price. By using related information, we can supplement the information obtained, broaden the analytical perspective, and make the subsequent predicted price more objective and accurate.

[0046] The associated information includes spatial geographic information. Spatial geographic information is used to characterize the spatial features of the target property within a preset range. This can be image-based information, such as images, within the preset range. Specifically, spatial geographic information includes the location distribution of the target property within the preset range, its spatial relationship with other elements (such as other properties and facilities), and its geographic features. The preset range is a pre-defined spatial area. For example, the preset range can be a circular area centered on the target property with a preset distance as its radius; or, it can be the spatial area of ​​the street, district, or city where the target property is located.

[0047] Based on spatial geographic information, data such as the building's exterior materials, window ratio, special facilities (e.g., swimming pools), greening rate, spatial relationship with infrastructure (e.g., high-voltage power lines), and building density within a preset range can be extracted for the target real estate. In one example, if the real estate is land, information on crops on the land can also be extracted based on spatial geographic information to determine the economic value of the crops.

[0048] Spatial geographic information can be acquired through technologies such as satellite imagery, drone aerial photography, or mapping software to obtain data (such as image data) within a predetermined area of ​​the target real estate, thereby generating spatial geographic information.

[0049] Step S102: Generate a feature matrix based on attribute information and association information.

[0050] Specifically, a first feature vector is generated based on the attribute information; a second feature vector is generated based on the association information; and the first and second feature vectors are concatenated to obtain a feature matrix.

[0051] Generating a first feature vector based on attribute information can be achieved by quantifying each indicator in the attribute information and then defining each quantified indicator as an element of the first feature vector. For example, for numerical indicators such as area and building age, the magnitude of each indicator can be standardized, converting the numerical indicators into values ​​of the same magnitude; for categorical indicators such as structure and usage, corresponding values ​​for each type can be pre-set; and for textual indicators such as location, key information can be extracted, converting the textual semantics into numerical values.

[0052] The first feature vector is generated based on the attribute information. Alternatively, a base price for the target real estate can be determined based on the attribute information; this base price is the first feature vector. This base price is a price predicted solely based on the attribute information of the target real estate itself.

[0053] For example, the base price of the target property can be determined based on its attribute information and market transaction data (such as historical transaction records and auction data). For instance, based on the target property's area, location, and other attribute information, similar cases (properties in the same neighborhood or with the same unit type) can be identified from market transaction data, and the transaction prices corresponding to these similar cases can be used as the base price.

[0054] In this embodiment, a second feature vector is generated based on the associated information, specifically based on spatial geographic information.

[0055] Based on spatial geographic information, a second feature vector is generated. This can be achieved by identifying elements within the spatial geographic information, determining multiple indicators based on these elements, quantifying each indicator, and then defining the quantified indicators as elements of the second feature vector. These elements include relevant infrastructure, special facilities, and buildings within the spatial geographic information.

[0056] For example, the external features of the building housing the target property in the spatial image are extracted, and the external material of the building is determined by combining the typical characteristics of different materials (such as texture and color). Green areas in the spatial image are identified and their areas are calculated; the greening rate is determined based on the ratio of the green area to the area of ​​a preset range. The locations of infrastructure facilities in the spatial image are identified, and the spatial relationship between the target property and the infrastructure facilities is calculated, such as the distance between the infrastructure facilities and the target property. The number of buildings in the spatial image is identified, and the building density within the preset range is determined based on the ratio of the number of buildings to the area of ​​the preset range.

[0057] The external materials, greening rate, spatial relationship with infrastructure, and building density are quantified to obtain the elements of the second feature vector.

[0058] Based on spatial geographic information, a second feature vector can be generated. Alternatively, based on spatial geographic information, the influence factors of the environment on the price of the target real estate can be determined, and these influence factors can be used as the second feature vector.

[0059] For example, target properties with a higher greening rate in their surroundings tend to have higher prices, which can correspondingly increase the impact factor; target properties with a higher building density in their surroundings may also have higher prices, which can also correspondingly increase the impact factor.

[0060] Step S103: Determine the target valuation model from multiple pre-trained valuation models based on the type of the target real estate.

[0061] The types of real estate include, but are not limited to, residential real estate used solely to meet residential needs, residential real estate with special value (such as school district housing, historical buildings, etc.), commercial real estate (such as shops), industrial real estate (such as factories), and real estate for special purposes (such as land, homesteads, etc.).

[0062] Valuation models can be statistical or machine learning models, such as multiple linear regression, random forest, or neural network models. Valuation models can analyze and learn from input features to determine the relationship between those features and real estate prices.

[0063] Specifically, the valuation model is trained using feature matrices of multiple real estate properties as a training set and their prices as labels. The property prices can be either predicted prices or transaction prices. Through training, the parameters of the valuation model are adjusted so that it learns the relationship between the feature matrices and the prices. The trained valuation model can then output an accurate predicted price based on the input feature matrix of the target real estate property.

[0064] The pricing logic of different types of real estate differs; that is, the price drivers and market rules for residential, commercial, and industrial real estate are completely different. Using the same calculation method cannot be fully applied to all types of real estate. Simply using a single valuation model to determine the predicted price of all types of real estate may result in low prediction accuracy and weak reference value of the prediction results.

[0065] To address this issue, this application pre-trains multiple valuation models corresponding to different types of real estate. Specifically, the parameters of the corresponding valuation models are adjusted based on the price composition logic of different types of real estate.

[0066] For example, for residential properties with special value, a special value weighting coefficient can be added; similarly, for properties with historical architectural value, a cultural value weighting coefficient can be added to increase the focus on special value. For industrial properties or special-purpose properties (such as land), an economic value weighting coefficient can be added to increase the focus on the economic value of the goods or crops produced.

[0067] In some embodiments, multiple valuation models can be used, employing different technologies based on the complexity of the constituent logic of the corresponding type of real estate. For example, for real estate with simple constituent logic, such as residential real estate used solely to meet living needs, a multiple linear regression model is used; for real estate with complex constituent logic, a random forest model or a neural network model is used.

[0068] By pre-training multiple valuation models, when it is necessary to predict the price of a target real estate, the target valuation model corresponding to that type is determined from the multiple pre-trained valuation models according to the type of the target real estate.

[0069] Step S104: Input the feature matrix into the target valuation model to obtain the predicted price of the target real estate.

[0070] After determining the target valuation model, the feature matrix is ​​input into the target valuation model. The target valuation model analyzes the feature matrix and outputs the predicted price of the target real estate.

[0071] Due to the model's powerful computing capabilities, the time to obtain predicted prices is generally in the seconds, compared to the 3-7 days required for manual assessment. This method significantly saves time costs and also saves financial costs because it does not require on-site investigation. At the same time, the model's standardized and objective nature avoids the subjective biases caused by manual assessment.

[0072] The real estate price prediction method provided in this embodiment acquires the property's own attribute information and related information to evaluate the property from a more comprehensive perspective. Then, based on the acquired information, a feature matrix representing the property is generated, and a corresponding model is determined from multiple pre-trained models. By selecting the matching model, the model's ability to capture specific data patterns and relationships can be fully utilized, thereby improving prediction accuracy. Finally, the feature matrix is ​​input into the model, and through the model's powerful learning and reasoning capabilities, the model automatically outputs the predicted price of the real estate. By outputting the prediction results through the model, the time required for obtaining predicted prices through manual evaluation is significantly reduced, improving prediction efficiency. Furthermore, by evaluating the real estate from a more comprehensive perspective, the final prediction results are more accurate.

[0073] In one possible implementation, the associated information includes spatial geographic information.

[0074] Step S102: Generate a feature matrix based on attribute information and association information, including: extracting target element information from spatial geographic information in association information; and generating a feature matrix based on attribute information and target element information.

[0075] Target elements are the relevant infrastructure, special facilities, buildings, etc., that need to be analyzed in spatial geographic information. Common target elements include schools, hospitals, residences, shops, transportation hubs, green areas, etc. Information about target elements includes, but is not limited to, their geographical location, quantity, and density.

[0076] Spatial geographic information can be image-based, such as spatial images. When spatial geographic information is a spatial image, image recognition techniques, such as deep learning image recognition, are used to identify information about target elements within the spatial geographic information, such as the geographical location, quantity, and density of the target elements. Alternatively, spatial geographic information can also be vector-based, such as the coordinate information of elements. When spatial geographic information is coordinate information, the coordinates of each target element are extracted, and based on the coordinates of the target elements, the information about the target elements, such as their geographical location, quantity, and density, is determined. Then, based on the attribute information, a first feature vector is determined; based on the information of the target elements, a second feature vector is determined; the first and second feature vectors are concatenated to obtain a feature matrix.

[0077] Based on the information of the target element, the second feature vector is determined. This can be achieved by quantifying the information of the target element to obtain the elements of the second feature vector.

[0078] Optionally, the spatial geographic information includes map data within a preset range of the target real estate, and the method further includes: determining the path between the target real estate and the target building based on the map data within the preset range of the target real estate; and determining the target element from the elements along the path.

[0079] The target building is a frequently visited building in real-life scenarios, such as a transportation hub.

[0080] Within the map data of the target property's pre-defined area, determine the actual path between the target property and the target building, rather than the straight-line connection between them. This path is typically the route frequently taken by residents of the target property. Identify the target element from all elements traversed by this path.

[0081] For example, Figure 2 This is a schematic diagram of map data within a preset range provided in this application. For example... Figure 2 As shown, the map data is the map data of the street where the target real estate is located, and the preset range is the street. Figure 2With the subway station as the target building, the path indicated by the arrow represents the route between the target property and the target building. According to... Figure 2 It can be seen that the elements traversed by this path include: Shop 1, Shop 2, Shop 5, Residential 1, Residential 3, and Residential 5. When information about residential properties is needed, the target elements are Residential 1, Residential 3, and Residential 5; when information about commercial properties is needed, the target elements are Shop 1, Shop 2, and Shop 5.

[0082] Based on the target element information in spatial geographic information, a feature matrix is ​​generated, enabling precise mining and analysis of elements surrounding the target real estate. This provides a more comprehensive and objective basis for prediction, thereby enhancing the accuracy and reliability of the prediction results. Simultaneously, by determining the path between the target real estate and the target building, target elements are identified. These identified target elements have a stronger correlation with the target real estate and a greater impact on its price, achieving more precise target element positioning.

[0083] Figure 3 This is a flowchart illustrating another method for predicting real estate prices provided in this embodiment of the application. The method for predicting real estate prices provided in this embodiment is... Figure 1 Based on the illustrated embodiment, steps S101 and S102 are refined, and a model retraining step is added. For example... Figure 3 As shown, the method provided in this embodiment includes the following steps:

[0084] Step S301: Obtain the attribute information and related information of the target real estate.

[0085] The associated information includes spatial geographic information and regional fluctuation information. The regional fluctuation information is used to describe the fluctuation trend of preset parameters in the area where the target real estate is located.

[0086] Regional fluctuation information includes supporting data for the target real estate in the city or region, such as the number of schools, subways and shopping malls, environmental data (such as air quality), population inflow trends, GDP of the city or region, and regional policies (such as housing purchase policies and regional redevelopment policies).

[0087] Information on regional fluctuations can be obtained in real time through the internet by accessing information released on official websites or by media outlets.

[0088] Step S302: Based on natural language processing technology, extract target information related to the target real estate from the regional fluctuation information.

[0089] Regional fluctuation information is a text-based information. To facilitate computer understanding, this embodiment uses natural language processing techniques, such as general pre-trained models based on BERT (Bidirectional Encoder Representations from Transformers) and RoBERTa (Robustly Optimized BERT Pretraining Approach), to extract target information related to the target real estate from the regional fluctuation information. Target information includes multiple items such as the number of schools, subways, and shopping malls, environmental data, population inflow trends, the GDP of the city or region, and regional policies (e.g., housing purchase policies, regional redevelopment policies).

[0090] Step S303: If the target information meets the preset premium conditions, then determine the premium factor of the target real estate.

[0091] A premium condition is a condition that causes the actual price of the target property to be higher than its benchmark price or market average price. This includes conditions such as the existence of redevelopment (e.g., demolition) policies in the area where the target property is located, a surge in urban population inflow, and the fact that the area where the target property is located is a school district or a low-lying area.

[0092] If the target information meets the preset premium conditions, the actual price of the target real estate may be higher than the benchmark price. Based on the target information, the premium factor of the target real estate is determined.

[0093] The premium factor can be determined by selecting the premium factor of the target real estate from the premium factors corresponding to the pre-set premium conditions, based on the premium conditions satisfied by the target information.

[0094] For example, the premium condition includes the existence of redevelopment policies in the area where the property is located. If the target information meets the premium condition, the premium factor of the target property is determined to be the premium factor corresponding to the premium condition.

[0095] Step S304: Generate a feature matrix based on attribute information, association information, and premium factor.

[0096] In this step, the first feature vector is obtained based on the attribute information; the second feature vector is obtained based on the spatial geographic information in the associated information; the premium factor is used as the third feature vector; and the first, second, and third feature vectors are concatenated to obtain the feature matrix.

[0097] In some embodiments, a fourth feature vector can be generated based on regional fluctuation information, and the first, second, third, and fourth feature vectors can be concatenated to obtain a feature matrix.

[0098] Based on regional fluctuation information, a fourth feature vector is generated. This can be used to extract information from regional fluctuation information that has a lasting impact on the price of target real estate, such as whether it is a school district or the region's GDP. This information is then quantified to obtain the elements of the fourth feature vector.

[0099] Step S305: Determine the target valuation model from multiple pre-trained valuation models based on the type of the target real estate.

[0100] The target real estate includes two types: Category I real estate and Category II real estate. Category I real estate refers to real estate whose building complex has a floor area ratio greater than or equal to a first threshold and whose geographical location is not within the target zoning area. Category II real estate refers to real estate whose building complex has a floor area ratio less than a second threshold or whose geographical location is within the target zoning area.

[0101] The target valuation model for the first type of real estate is a pre-trained first valuation model, which is constructed based on a multiple linear regression model; the target valuation model for the second type of real estate is a pre-trained second valuation model, which is constructed based on a random forest model.

[0102] The target school district can be a school zone. A building complex is a group of multiple independent buildings within the same regulated land area. The plot ratio is the ratio of the total building area of ​​a building complex within a predetermined area to the total land area within that area, reflecting land use efficiency.

[0103] In actual land planning, Category I real estate, where the plot ratio of a building complex is greater than or equal to the first threshold and its geographical location is not within the target zoning area, is typically residential real estate used solely to meet residential needs. Category II real estate, where the plot ratio of a building complex is less than the second threshold, or its geographical location is within the target zoning area, is typically a villa area or school district housing. The first and second thresholds are parameters determined according to land planning.

[0104] Multiple linear regression is a classic predictive model in statistics and machine learning used to analyze the linear relationship between multiple independent variables (elements in a feature matrix) and a dependent variable (predicted price). Its core idea is to construct a linear equation by quantifying the influence of independent variables on the dependent variable to predict the dependent variable.

[0105] Random forest is a supervised learning algorithm based on ensemble learning. It improves model performance by constructing multiple decision trees and combining the predictions of all trees.

[0106] Since the price structure of Category I real estate is relatively simple and typically has no other special factors affecting the price, a simple model can be used to accurately predict the price. For Category I real estate, a first valuation model based on a multiple linear regression model is chosen to predict its price. Since the price structure of Category II real estate is more complex, a second valuation model based on a random forest model is chosen to predict its price.

[0107] Step S306: Input the feature matrix into the target valuation model to obtain the predicted price of the target real estate.

[0108] Optionally, if the target property is of type 'target type', which is a type other than Class I and Class II real estate, the method further includes: using a pre-trained neural network model to obtain the predicted price of the target property; or, generating manual prompts to instruct relevant personnel to manually predict the price of the target property.

[0109] Besides Category I and Category II real estate, real estate also includes other special types, such as disputed real estate, whose price composition logic is more complex. For this type of real estate, a pre-trained neural network model can be used to predict the price of the target real estate. Alternatively, human-generated prompts can be directly generated to allow for manual evaluation of complex target real estate, i.e., relevant personnel can manually predict the price of the target real estate.

[0110] By using neural network models and manual assessments, the price prediction method for this target real estate becomes more flexible, enabling targeted price predictions and more accurate results.

[0111] Step S307: Obtain the transaction price of the newly transacted target real estate.

[0112] In this step, the transaction price of the target real estate in the market transaction data is obtained according to a preset period. In this embodiment, the target real estate can be real estate whose value has been predicted using the real estate value prediction method of this application.

[0113] Step S308: If the difference between the transaction price and the predicted price meets the preset conditions, then the target valuation model is retrained based on the new sample.

[0114] If the deviation between the transaction price and the predicted price is large, that is, if the difference between the transaction price and the predicted price meets the preset condition, such as being greater than or equal to 15% of the predicted price, it indicates that the accuracy of the target valuation model is low, and the target valuation model will be retrained.

[0115] Specifically, the feature matrix of the newly transacted target real estate is used as a new sample, and the transaction price is used as the label of the new sample to retrain the target valuation model.

[0116] In this embodiment, by adding regional fluctuation information, the information dimensions of the method are expanded, providing a more comprehensive perspective and thus making the subsequently obtained predicted prices more accurate. Furthermore, by determining the premium factor based on regional fluctuation information, the regional fluctuation information is quantified, avoiding excessive resource consumption caused by the model analyzing a large amount of regional fluctuation data. Simultaneously, by distinguishing the characteristics of the first and second types of real estate, it is easier for the computer to identify the type of target real estate and select different models for prediction based on the type of real estate, resulting in more accurate predictions. In addition, by retraining the model when the accuracy of the target valuation model is low, the timeliness and accuracy of the model are ensured.

[0117] In one possible implementation, since the prices of real estate in the same area or community are similar, if a real estate in the same area or community has a predicted price obtained through the real estate price prediction method provided in this application, then the predicted price of that real estate is applicable to other real estate in the same area or community.

[0118] This application also provides a method for determining the loan amount. The method includes determining the loan amount based on the predicted price of the mortgaged real estate of the loan user.

[0119] The predicted price is the predicted price obtained by the real estate price prediction method provided in the above embodiments.

[0120] The method for determining the loan amount provided in this application can be implemented by the respective banking and financial systems.

[0121] Specifically, the banking system obtains the predicted price of the real estate using the real estate price prediction method provided in the above embodiments; based on the predicted price, the loan amount is obtained through the loan amount calculation method within the banking system. The banking system can then issue loans to loan users based on this loan amount.

[0122] The calculation method for loan amounts is the internal calculation method of each bank's financial system and is not specified here.

[0123] By using accurate and efficient methods for predicting real estate prices, the efficiency of loan amount determination can be improved, and the loan amount can be made more accurate, making the loan business process within the banking and financial system smoother and more reliable.

[0124] In one possible implementation, the banking and financial system can automatically capture regional fluctuation information. If relevant policy adjustments are detected, such as changes in school district policies, changes in transportation hub planning, or urban redevelopment plans, the real estate price prediction method and the loan amount determination method provided in the above embodiments are executed to dynamically update the predicted price of the mortgaged real estate and the loan amount.

[0125] Figure 4 This is a schematic diagram of a real estate price prediction device provided in an embodiment of this application. Figure 4 As shown, the real estate price prediction device provided in this embodiment includes an information acquisition module 401, a feature matrix generation module 402, a model determination module 403, and a price prediction module 404.

[0126] The information acquisition module 401 is used to acquire the attribute information and related information of the target real estate, including spatial geographic information; the spatial geographic information is used to characterize the spatial information of the target real estate within a preset range; the feature matrix generation module 402 is used to generate a feature matrix based on the attribute information and related information; the model determination module 403 is used to determine the target valuation model from multiple pre-trained valuation models based on the type of the target real estate; the price prediction module 404 is used to input the feature matrix into the target valuation model to obtain the predicted price of the target real estate.

[0127] Optionally, the feature matrix generation module 402 is specifically used for:

[0128] Extract target element information from spatial geographic information in the associated information; generate a feature matrix based on attribute information and target element information.

[0129] Optionally, the spatial geographic information includes map data within a preset area of ​​the target real estate. The feature matrix generation module 402 is also used for:

[0130] Based on map data within the preset area of ​​the target real estate, determine the path between the target real estate and the target building; identify the target element from the elements along the path.

[0131] Optionally, the associated information may also include regional fluctuation information, which describes the fluctuation trend of preset parameters in the region where the target real estate is located.

[0132] Optionally, the feature matrix generation module 402 is specifically used for:

[0133] Based on natural language processing technology, target information related to the target real estate is extracted from regional fluctuation information; if the target information meets the preset premium conditions, the premium factor of the target real estate is determined; and a feature matrix is ​​generated based on attribute information, correlation information and premium factor.

[0134] Optionally, the target real estate includes two types: Type I real estate and Type II real estate. Type I real estate refers to real estate whose building density and floor area ratio are greater than or equal to a first threshold and whose geographical location is not within the target zoning area. Type II real estate refers to real estate whose building density and floor area ratio are less than a second threshold, or whose geographical location is within the target zoning area. The target valuation model corresponding to Type I real estate is a pre-trained first valuation model, which is constructed based on a multiple linear regression model. The target valuation model corresponding to Type II real estate is a pre-trained second valuation model, which is constructed based on a random forest model.

[0135] Optionally, the real estate price forecasting device further includes a second forecasting module, which is used for:

[0136] If the target property is of type Target Type, which is a type other than Type I and Type II real estate, a pre-trained neural network model is used to obtain the predicted price of the target property; or, manual prompts are generated to instruct relevant personnel to predict the price of the target property manually.

[0137] Optionally, the real estate price prediction device also includes a model optimization module, which is used for:

[0138] Obtain the transaction price of the newly transacted target real estate; if the difference between the transaction price and the predicted price meets the preset conditions, retrain the model based on the new sample.

[0139] The real estate price prediction device provided in this application can be used to execute the real estate price prediction method provided in any of the above embodiments of this application. Its implementation principle and technical effect are similar, and will not be described again here.

[0140] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 5 As shown, the electronic device of this embodiment may include: at least one processor 501; and a memory 502 communicatively connected to the at least one processor; wherein the memory 502 stores instructions executable by the at least one processor 501, the instructions being executed by the at least one processor 501 to cause the electronic device to perform the method as described in any of the above embodiments.

[0141] Optionally, the memory 502 can be either standalone or integrated with the processor 501. When the memory 502 is set up independently, the device also includes a bus for connecting the memory 502 and the processor 501.

[0142] The implementation principle and technical effects of the electronic device provided in this embodiment can be found in the foregoing embodiments, and will not be repeated here.

[0143] This application also provides a computer-readable storage medium storing computer-executable instructions. When the computer-executable instructions are executed by a processor, the methods provided in any of the foregoing embodiments can be implemented.

[0144] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the method provided in any of the foregoing embodiments.

[0145] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0146] It should be further noted that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0147] It should be understood that the above-described device embodiments are merely illustrative, and the device of this application can also be implemented in other ways. For example, the division of units / modules in the above embodiments is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units, modules, or components may be combined, or integrated into another system, or some features may be ignored or not executed.

[0148] Furthermore, unless otherwise specified, the functional units / modules in the various embodiments of this application can be integrated into one unit / module, or each unit / module can exist physically separately, or two or more units / modules can be integrated together. The integrated units / modules described above can be implemented in hardware or as software program modules.

[0149] When integrated units / modules are implemented in hardware, the hardware can be digital circuits, analog circuits, etc. The physical implementation of the hardware structure includes, but is not limited to, transistors, memristors, etc. Unless otherwise specified, the processor can be any suitable hardware processor, such as a CPU, GPU, FPGA, DSP, and ASIC, etc. Unless otherwise specified, the storage unit can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc.

[0150] If the integrated unit / module is implemented as a software program module and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0151] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0152] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0153] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for predicting the price of real estate, characterized in that, include: Obtain the attribute information of the target real estate and the associated information of the target real estate, wherein the associated information includes spatial geographic information; The spatial geographic information is used to characterize the spatial features of the target real estate within a preset range; Generate a feature matrix based on the attribute information and the association information; Based on the type of the target real estate, a target valuation model is determined from multiple pre-trained valuation models; The feature matrix is ​​input into the target valuation model to obtain the predicted price of the target real estate.

2. The method according to claim 1, characterized in that, The step of generating a feature matrix based on the attribute information and the association information includes: Extract information about the target element from the spatial geographic information in the associated information; The feature matrix is ​​generated based on the attribute information and the target element information.

3. The method according to claim 2, characterized in that, Spatial geographic information includes map data within a preset area of ​​the target real estate, and the method further includes: Based on map data within a preset area of ​​the target real estate, determine the path between the target real estate and the target building; The target element is determined from the elements traversed by the path.

4. The method according to claim 1, characterized in that, The associated information also includes regional fluctuation information, which is used to describe the fluctuation trend of preset parameters in the region where the target real estate is located.

5. The method according to claim 4, characterized in that, The step of generating a feature matrix based on the attribute information and the association information includes: Based on natural language processing technology, target information related to the target real estate is extracted from the regional fluctuation information; If the target information meets the preset premium conditions, then the premium factor of the target real estate is determined; The feature matrix is ​​generated based on the attribute information, the association information, and the premium factor.

6. The method according to claim 1, characterized in that, The target real estate includes a first type of real estate and a second type of real estate; wherein, the first type of real estate is real estate whose building complex has a floor area ratio greater than or equal to a first threshold and whose geographical location is not within the target zoning area; the second type of real estate is real estate whose building complex has a floor area ratio less than a second threshold, or whose geographical location is within the target zoning area. The target valuation model for the first type of real estate is a pre-trained first valuation model, which is constructed based on a multiple linear regression model; the target valuation model for the second type of real estate is a pre-trained second valuation model, which is constructed based on a random forest model.

7. The method according to claim 6, characterized in that, If the target real estate is of a target type, and the target type is a type other than the first type of real estate and the second type of real estate, the method further includes: The predicted price of the target real estate is obtained using a pre-trained neural network model; or, a human prompt is generated to instruct relevant personnel to predict the price of the target real estate manually.

8. The method according to any one of claims 1-7, characterized in that, After generating the predicted price, the method further includes: Obtain the transaction price of the newly sold target real estate; If the difference between the transaction price and the predicted price meets a preset condition, the target valuation model is retrained based on the new sample.

9. A method for determining loan amount, characterized in that, include: The loan amount is determined based on the predicted price of the mortgaged real estate of the loan user; wherein the predicted price is the predicted price obtained by the method according to any one of claims 1-8.

10. A real estate price prediction device, characterized in that, include: The information acquisition module is used to acquire the attribute information of the target real estate and the associated information of the target real estate, wherein the associated information includes spatial geographic information; The spatial geographic information is used to characterize the spatial information of the target real estate within a preset range; The feature matrix generation module is used to generate a feature matrix based on the attribute information and the association information; The model determination module is used to determine a target valuation model from multiple pre-trained valuation models based on the type of the target real estate. The price prediction module is used to input the feature matrix into the target valuation model to obtain the predicted price of the target real estate.

11. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1 to 9.

13. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 9.

Images