Method for measuring and calculating sparse sample anchor point compensation value

Abstract

A sparse sample anchor point compensation value measuring and calculating method comprises a data acquisition module, a value measuring and calculating module and a comprehensive measuring and calculating module, and the data acquisition module comprises a commodity residence data acquisition module C1, an apartment market data acquisition module F1, an apartment product feature acquisition module F2 and a residence and apartment periphery matching acquisition module Z1; the value measuring and calculating module comprises a commercial residence rental price measuring and calculating module PC and a promotion factor measuring and calculating module PF; the commercial residence rental price measuring and calculating module PC obtains a basic price P0 through model regression analysis; the promotion factor measurement and calculation module PF performs processing through feature engineering according to the apartment market data acquisition module F1, the apartment product feature acquisition module F2 and the residence and apartment periphery matching acquisition module Z1 to obtain an apartment market promotion factor X1 and an apartment house type matching promotion factor X2; and the comprehensive calculation module 3 accurately estimates the optimal rent of the apartment on the market through P = P0 (1 + X1) * (1 + X2).

Description

Technical Field

[0001] The present invention relates to the technical field of land price calculation, and in particular to a method for calculating the compensation value of sparse sample anchor points. Background Art

[0002] Real estate refers to objects that occupies a fixed position in space due to natural properties or legal regulations and whose economic value will be affected by moving. It includes land, land fixtures, land products that have not yet been separated from the land, and other objects that are attached to the land by nature or human power and cannot be separated.

[0003] At present, the geographical location, price and environment of the house are recorded in the registration of real estate. When users want to sell or rent a house in real estate, they need to know the price and rental price of the same type of house in advance. There are two common ways for users to learn about house prices. One is to inquire through an intermediary agency; the other is that users need to search for relevant information online and organize the relevant information. Due to the rapid development of the Internet, when users search for information about houses online, they cannot distinguish the authenticity of the information about houses online, resulting in users spending a lot of time to organize the house prices that are inaccurate, which in turn affects users' judgment of real estate house prices.

[0004] The existing processing method is mainly based on the location competitive product comparison method, finding the same type or similar samples within a certain radius of the surrounding area, and realizing price estimation by comparing the advantages and disadvantages. Here, the basic price is evaluated based on the similarity of the location and then the heterogeneity is adjusted; a method and device for predicting the operating price of real estate (CN202310418441.2) emphasizes the use of computers to predict real estate prices. The above patent does not solve the rental, cash flow or value prediction of real estate in the case of sparse samples or missing samples. There is room for improvement in the existing technology. Summary of the invention

[0005] In order to solve the above technical problems, the present invention proposes a method for calculating the compensation value of sparse sample anchor points. The price of commercial housing is used as a reference value, and the improvement factor and apartment product characteristics are combined to correct the rental price of apartment housing, so as to achieve quantitative and accurate estimation of apartment prices.

[0006] To achieve the above object, the technical solution of the present invention is as follows: a method for calculating the compensation value of sparse sample anchor points, comprising a data acquisition module, a value calculation module and a comprehensive calculation module; wherein:

[0007] The data collection module includes a commercial housing data collection module C1, an apartment market data collection module F1, an apartment product feature collection module F2, and a residential and apartment surrounding supporting collection module Z1;

[0008] The value calculation module includes: a commodity housing rental price calculation module PC and a boost factor calculation module PF; the commodity housing rental price calculation module PC includes a computer and a machine learning model, and processes the commodity housing data collection module C1 and the residential and apartment surrounding supporting collection module Z1 through feature engineering to extract features Y1 (population density), Y2 (transportation supporting facilities), Y3 (medical supporting facilities), Y4 (residential rental price), etc. The machine learning model obtains the basic price P0 through model regression analysis; the apartment feature calculation module PF processes the apartment market data collection module F1 through feature engineering to extract features U1 (area), U2 (decoration level), etc. The boost factor calculation module PF includes a computer and a machine learning model, and the boost factor calculation module PF obtains the apartment market boost factor X1 and the apartment room supporting boost factor X2 according to the apartment market data collection module F1, the apartment product feature collection module F2 and the residential and apartment surrounding supporting collection module Z1 through feature engineering;

[0009] The comprehensive calculation module 3 calculates the final price through P=P0(1+X1)*(1+X2).

[0010] By adopting the above technical scheme, the basic price P0 of the commodity housing rental price calculation module is estimated according to the characteristics collected by the commodity housing data collection module, and the housing improvement factor X1 and the apartment improvement factor X2 are obtained through the data collected by the apartment market data collection module F1, the apartment product feature collection module F2 and the apartment surrounding supporting collection module Z1. Through P=P0(1+X1)*(1+X2), the optimal rent of the apartment in the market is accurately estimated according to the prices of commodity housing around the apartment.

[0011] The present invention is further configured as follows: the commercial housing data acquisition module C1 acquires image information and text data, wherein the image information at least includes commercial housing photos, commercial housing floor plans, and commercial housing decoration levels; the text data at least includes commercial housing source information, listing information, transaction information, room types and layouts, rental information, and community information.

[0012] The present invention is further configured as follows: the apartment market data collection module F1 collects information including at least apartment type information, apartment room type information, apartment listing information and apartment transaction information.

[0013] The present invention is further configured as follows: the apartment product feature collection module F2 collects information including at least service type, decoration level, functional design, and internal configuration.

[0014] The present invention is further configured as follows: the residential and apartment surrounding supporting facilities collection module Z1 collects information including at least educational facilities, commercial facilities, medical facilities, and transportation facilities within a distance of 1-10 km around the commercial residential or apartment.

[0015] The present invention is further configured as follows: the improvement factor calculation module PF performs parameter fitting and optimization by iteratively minimizing the loss function to complete the calculation of the apartment market improvement factor X1 and the apartment room type supporting improvement factor X2.

[0016] In summary, the present invention has the following effects:

[0017] 1. The rental amount can be adjusted according to the educational facilities, commercial facilities, medical facilities and transportation facilities near the apartment;

[0018] 2. Ability to modify the lease amount based on the age of the community, community type, property conditions and land type;

[0019] 3. Ability to modify the rental amount by image recognition and decoration classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below.

[0021] Figure 1 It is the relationship diagram of data collection module, value calculation module and comprehensive calculation module;

[0022] Figure 2 It is a schematic diagram of information collection by the data collection module;

[0023] Figure 3 Schematic diagram of the machine learning model in the value measurement module. DETAILED DESCRIPTION

[0024] The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0025] like Figure 1 As shown, a method for calculating the compensation value of sparse sample anchor points includes a data acquisition module, a value calculation module and a comprehensive calculation module; wherein:

[0026] The data collection module includes a commercial housing data collection module C1, an apartment market data collection module F1, an apartment product feature collection module F2, and a residential and apartment surrounding supporting collection module Z1;

[0027] The commercial housing data collection module C1 collects image information and text data, wherein the image information at least includes commercial housing photos, commercial housing floor plans, and commercial housing decoration levels; the text data at least includes commercial housing source information, listing information, transaction information, room types and community information.

[0028] The apartment product feature collection module F2 collects information including at least service type, decoration level, functional design, and internal configuration.

[0029] The apartment market data collection module F1 collects information including at least apartment type information, apartment room type information, apartment listing information and apartment transaction information.

[0030] The value calculation module includes: a commodity housing rental price calculation module PC and a boost factor calculation module PF; the commodity housing rental price calculation module PC includes a computer and a machine learning model, processes the commodity housing data collection module C1 through feature engineering, extracts features Y1 (population density), Y2 (transportation facilities), Y3 (medical facilities), Y4 (housing rental price), etc., and the machine learning model obtains the basic price P0 through model regression analysis; the apartment feature calculation module PF processes the apartment market data collection module F1 through feature engineering, extracts features U1 (area), U2 (decoration level), etc., and the boost factor calculation module PF includes a computer and a machine learning model, and the boost factor calculation module PF obtains the apartment market boost factor X1 and the apartment room type supporting boost factor X2 according to the apartment market data collection module F1, the apartment product feature collection module F2, and the residential and apartment surrounding supporting collection module Z1;

[0031] The comprehensive calculation module 3 calculates the final price through P=P0(1+X1)*(1+X2).

[0032] The residential and apartment surrounding supporting facilities collection module Z1 collects information including at least educational facilities, commercial facilities, medical facilities, and transportation facilities within a distance of 1-10 km around the commercial residential or apartment.

[0033] Furthermore, the improvement factor calculation module PF achieves parameter fitting and optimization by iteratively minimizing the loss function to complete the calculation of the apartment market improvement factor X1 and the apartment room supporting improvement factor X2.

[0034] like Figure 2As shown in the figure, the four acquisition modules are commercial housing data acquisition module C1, apartment market data acquisition module F1, apartment product feature acquisition module F2, and residential and apartment surrounding supporting acquisition module Z1. Commercial housing data acquisition module C1 collects image and text data, which are personal housing information, listing information, and transaction information. Housing information includes photos of housing, room type diagrams, room type and apartment type, community information, etc. The surrounding supporting acquisition module collects geographic information, including educational supporting facilities, commercial supporting facilities, medical supporting facilities, and transportation supporting facilities. The apartment market data acquisition module F1 collects apartment type information, room type information, listing information, and transaction information. The apartment product feature acquisition module collects service type, room type, decoration level, area, functional design, and internal configuration. The collected information enters the second-level measurement module.

[0035] The second layer modules include commodity housing rental price calculation module PC and promotion factor calculation module PF. Figure 3 As shown, the commodity housing rental price calculation module PC and the improvement factor calculation module PF calculate the rental price of personal housing at the location through machine learning models. The calculation factors include population, industry, commercial supporting facilities, medical supporting facilities, educational supporting facilities, transportation supporting facilities, and the rent of different types of personal housing in the same area and location. The personal housing and apartment price improvement factor calculation module uses the above-mentioned personal housing rental prices as a reference, and establishes the corresponding ratio between centralized and personal housing according to the different types of other locations in the city (including serviced and non-serviced, long-term rental apartments and affordable rental housing), different room types (one-bedroom, two-bedroom, three-bedroom, four-bedroom, etc., shared rental), and the measured commodity housing rental prices. The apartment product feature collection module F2 calculates the apartment price in any location through this ratio combined with service type, room type, decoration level, area, functional design, and internal configuration. The data of these two modules are aggregated to the comprehensive calculation module on the third layer.

[0036] The comprehensive calculation module will comprehensively calculate the rental results based on P0 obtained by the commercial housing rental price calculation module PC, the apartment market improvement factor X1 obtained by the improvement factor calculation module PF, and the apartment room supporting improvement factor X2.

[0037] like Figure 3The individual residential rental price calculation module shown will establish model relationships for multiple types of data through artificial intelligence and machine learning algorithms to form a commodity residential rental price calculation model and a lifting factor calculation model. These data include image data, text data, and geographic information data. The system will classify the image data to form multiple types of decoration levels. For text information, it will further identify and clean it into the age of the community, community type, property fees, and land type. For geographic information, it will include educational facilities, commercial facilities, medical facilities, and transportation facilities. Among them, educational facilities are divided into university education and primary and secondary education. Commercial facilities are divided into shopping malls and comprehensive shopping centers, convenience stores and vending machines, and restaurants. Medical facilities are divided into the number and distance of hospitals and the number and distance of tertiary hospitals. Transportation facilities are divided into the number and distance of subways, the number and distance of buses, the number and distance of high-speed rail stations, and the number and distance of airports.

[0038] Specific implementation process:

[0039] The first step is to calculate the price of a house in a certain community. First, the commercial housing data collection module C1 and the residential and apartment surrounding supporting collection module Z1 obtain surrounding real estate information through the Internet, including the surrounding similar house sales price, rental price, building area, house type and other information; then, the real estate decoration level is calculated as the basic feature through the supporting Internet pictures; then, the basic features are established through the community information, including property fees, years, types including relocation, old public housing, second-hand houses, and new houses; then, the basic features are established through land types, including residential land, commercial land, office land, comprehensive land, industrial land, education and scientific research land, etc. ; Then, basic characteristics are established through university education, including the number of colleges and universities within 10km and the size of students; Then, basic characteristics are established through primary and secondary schools, including the number of primary and secondary schools within 10km; Then, basic characteristics are established through surrounding commercial facilities, including the number of shopping malls and comprehensive shopping centers within 10km, the number of convenience stores and vending machines within 1km, and the number of restaurants within 1km; Finally, basic characteristics are established through surrounding medical facilities, including the number of medical institutions within 1km and the number of Class A hospitals within 2km. Then, basic characteristics are established through transportation construction, including but not limited to the number and coverage of public transportation facilities (such as subways and bus stops) within 3km, etc. The commodity housing rental price calculation module PC calculates that the monthly rent per square meter is 80 yuan.

[0040] The second step is to measure the comparative measurement of commercial housing and centralized housing of different types and room types. The price ratio of affordable rental housing (one-bedroom, two-bedroom, three-bedroom, four-bedroom, shared) to individual housing between the middle and outer ring roads in Pudong New Area is 1:0.92:0.95:0.94:1.5:0.98; the price ratio of long-term rental apartments to individual housing between the middle and outer ring roads in Pudong New Area is 1:0.96:0.94:0.9:1.8:0.85; the price ratio of serviced apartments (one-bedroom, two-bedroom, three-bedroom) to individual housing between the middle and outer ring roads in Pudong New Area is 1:0.9:0.85:0.6.

[0041] The third step is to improve the factor calculation module PF to collect the specific impact of service type, room type, decoration level, area function design, and internal configuration. For example, if the service is 7*24 hours butler service, the room type is a one-bedroom LOFT, the decoration is fine decoration, the area is 25 square meters, the functions include bedroom, bathroom, living room, study, kitchen, and the internal configuration includes bed, sofa, kitchen stove, integrated toilet, etc., then an additional gain of 12% can be obtained.

[0042] The fourth step comprehensive calculation module comprehensive calculation is 80 yuan*1 / 0.98*1.12.

[0043] It should be pointed out that a person skilled in the art can make several modifications and improvements without departing from the inventive concept of the present invention, and these modifications and improvements all fall within the protection scope of the present invention.

Claims

1. A method for calculating the compensation value of sparse sample anchor points, characterized in that: It includes data collection module, value calculation module and comprehensive calculation module; among which: The data collection module includes a commercial housing data collection module C1, an apartment market data collection module F1, an apartment product feature collection module F2, and a residential and apartment surrounding supporting collection module Z1; The value calculation module includes: a commodity housing rental price calculation module PC and a boost factor calculation module PF; the commodity housing rental price calculation module PC includes a computer and a machine learning model, and processes the commodity housing data collection module C1 and the residential and apartment surrounding supporting collection module Z1 through feature engineering to extract features Y1 (population density), Y2 (transportation supporting facilities), Y3 (medical supporting facilities), Y4 (residential rental price), etc. The machine learning model obtains the basic price P0 through model regression analysis; the apartment feature calculation module PF processes the apartment market data collection module F1 through feature engineering to extract features U1 (area), U2 (decoration level), etc. The boost factor calculation module PF includes a computer and a machine learning model, and the boost factor calculation module PF obtains the apartment market boost factor X1 and the apartment room supporting boost factor X2 according to the apartment market data collection module F1, the apartment product feature collection module F2 and the residential and apartment surrounding supporting collection module Z1 through feature engineering; The comprehensive calculation module 3 calculates the final price through P=P0(1+X1)*(1+X2).

2. A method for calculating the compensation value of sparse sample anchor points according to claim 1, characterized in that: The commercial housing data collection module C1 collects image information and text data, wherein the image information at least includes commercial housing photos, commercial housing floor plans, and commercial housing decoration levels; the text data at least includes commercial housing source information, listing information, transaction information, room types and layouts, rental information, and community information.

3. The method for calculating the compensation value of sparse sample anchor points according to claim 1, characterized in that: The apartment market data collection module F1 collects information including at least apartment type information, apartment room type information, apartment listing information and apartment transaction information.

4. The method for calculating the compensation value of sparse sample anchor points according to claim 3, characterized in that: The apartment product feature collection module F2 collects information including at least service type, decoration level, functional design, and internal configuration.

5. A method for calculating the compensation value of sparse sample anchor points according to claim 4, characterized in that: The residential and apartment surrounding supporting facilities collection module Z1 collects information including at least educational supporting facilities, commercial supporting facilities, medical supporting facilities, and transportation supporting facilities within a distance of 1-10 km around the commercial residential or apartment.

6. The method for calculating the compensation value of sparse sample anchor points according to claim 1, characterized in that: The improvement factor calculation module PF completes the calculation of the apartment market improvement factor X1 and the apartment room supporting improvement factor X2 by iteratively minimizing the loss function to achieve parameter fitting and optimization.

Images