Residence suitability index evaluation method and system fusing social perception spatio-temporal data

Abstract

The invention discloses a residential suitability index evaluation method and system fused with social perception spatio-temporal data, particularly relates to the technical field of urban planning, and aims to obtain related indexes of residential areas from a plurality of preset dimensions, and the residential areas include but are not limited to two types of urban village residences and commercial residences. The method comprises the following steps: performing feature extraction on each dimension, performing index acquisition evaluation based on a feature extraction result, then screening indexes of each dimension according to an index acquisition difficulty degree and a contribution degree to obtain an evaluation index set, and dividing the evaluation index set into a positive index set, a neutral index set and a negative index set according to a preset classification standard; converting the sets into corresponding index vectors through directivity conversion operation, inputting the index vectors into a pre-trained suitability evaluation model, and finally outputting a residence suitability score; according to the method, the complex relationship among the multi-dimensional data can be effectively processed, and the accurate suitability score is generated.

Description

Technical Field

[0001] The present invention relates to the field of urban planning technology, and more specifically, to a residential suitability index evaluation method and system integrating social perception spatiotemporal data. Background Art

[0002] With the rapid advancement of urbanization, the scale of cities continues to expand and the population density increases. How to reasonably plan urban space, optimize the living environment, and improve the quality of life of residents has become an important issue in the field of urban planning. Traditional urban planning is mostly based on physical factors such as land use, transportation network and infrastructure support. However, in today's complex urban environment, single-dimensional planning can no longer meet the diverse needs of residents. Therefore, in the process of urban planning, evaluating the suitability of residential areas has become a key means to improve the functional layout of cities and the happiness of residents.

[0003] Traditional residential suitability assessment is usually based on a single or a few indicators such as transportation convenience, land utilization rate, and infrastructure distribution. However, this assessment method has significant limitations because it cannot fully consider the comprehensive factors of the residential surrounding environment. For example, although transportation convenience can affect regional value, it ignores other important factors such as environmental pollution levels, accessibility of medical resources, and social security, which also have a profound impact on residents' living experience. In addition, traditional methods cannot flexibly respond to the diverse needs of residents in different areas, and the assessment results often cannot fully reflect the actual living conditions in different areas of the city.

[0004] With the development of big data, the Internet of Things, and smart city technologies, residential suitability assessment in urban planning has been able to obtain data sources in more dimensions, such as real-time environmental monitoring data collected by sensors, social perception data fed back by residents, etc. However, how to fully integrate and utilize these multi-dimensional data within the framework of urban planning, conduct a comprehensive and accurate assessment of residential suitability, improve the convenience and accuracy of real estate transactions, and provide accurate decision-making support for urban planners is still a challenge facing the current urban planning field. Summary of the invention

[0005] To achieve the above object, the present invention provides the following technical solutions: The residential suitability index evaluation method integrating social perception spatiotemporal data includes the following steps: Relevant indicators of residential areas are obtained from multiple preset dimensions respectively. Taking the preset dimensions as the target, the acquisition process of relevant indicators of each target dimension is subjected to feature extraction, and the difficulty of indicator acquisition and contribution are evaluated based on the feature extraction results; Screen the relevant indicators of the target dimension according to the difficulty of acquisition and contribution evaluation to obtain a set of evaluation indicators for the target dimension; Summarize the evaluation indicator sets of all target dimensions and divide them into positive indicator sets, neutral indicator sets, and negative indicator sets according to the preset classification standards; The positive indicator set, the neutral indicator set, and the negative indicator set are respectively subjected to directional conversion operations to obtain a positive index vector, a neutral index vector, and a negative index vector, and then the positive index vector, the neutral index vector, and the negative index vector are input together as input data into a pre-trained fitness evaluation model. The output result of the fitness evaluation model is the residential suitability.

[0006] In a preferred embodiment, the multiple preset dimensions refer to a social perception dimension, a spatiotemporal behavior dimension, a traffic convenience dimension, a medical resource dimension, a crisis response dimension, and an environmental pollution dimension.

[0007] In a preferred embodiment, a data indicator database is preset, which includes multiple groups of basic indicator sets, each basic indicator set corresponds to a target dimension, and the basic indicator set includes all relevant indicators of the corresponding target dimension.

[0008] In a preferred embodiment, taking the preset dimension as the target, when the process of acquiring relevant indicators of each target dimension performs feature extraction, the feature extraction includes acquisition difficulty information, computing cost information and linear related information.

[0009] In a preferred embodiment, evaluating the difficulty of obtaining indicators and their contribution based on the feature extraction results refers to: The difficulty evaluation value is generated based on the difficulty information obtained. The calculation formula is: ; is the difficulty evaluation value of the ith indicator, To obtain the time required for the i-th indicator, The economic cost of obtaining the i-th indicator is: The number of steps required to obtain the i-th indicator, is the availability coefficient of the ith indicator, , , These are all preset difficulty influence coefficients; Availability factor It is calculated by weighted summing the proportion of the area covered by the data source collection process corresponding to the i-th indicator to the entire residential area, the data source collection frequency, and the completeness of the data source; The cost evaluation value is generated based on the computing power cost information. The calculation formula is: ; is the cost evaluation value of the ith indicator, is the time required to calculate the i-th indicator, is the amount of data required for the i-th indicator, is the efficiency of parallel computing for the i-th indicator, , , All are preset cost impact coefficients; Based on the linear correlation information, a linear evaluation value is generated, and the Pearson correlation coefficient is used to evaluate the correlation between each indicator and residential suitability to obtain the correlation coefficient. , the linear evaluation value is marked as And the value and correlation coefficient The absolute values ​​are equal.

[0010] In a preferred embodiment, the relevant indicators of the target dimension are screened according to the difficulty of indicator acquisition and contribution evaluation, and the evaluation indicator set of the target dimension is obtained, which refers to: Compare the cost evaluation value of each indicator of the target dimension with a preset threshold value 1, eliminate indicators that are greater than or equal to the preset threshold value 1, and obtain an indicator set 1; Compare the difficulty evaluation value of each indicator of the target dimension with the preset threshold value 2, eliminate indicators that are greater than or equal to the preset threshold value 2, and obtain indicator set 2; Compare the linear evaluation value of each indicator of the target dimension with the preset threshold value three, eliminate the indicators that are less than the preset threshold value three, and obtain the indicator set three; The intersection of indicator set 1, indicator set 2 and indicator set 3 is obtained as the evaluation indicator set of the target dimension.

[0011] In a preferred embodiment, the directivity conversion operation refers to: For the forward indicator set, the following transformations are performed: ; A positive indicator The standard deviation of the time series corresponding to the time of collection, A positive indicator The average value of the time series corresponding to the time of collection, Indicates the offset set to prevent meaningless calculations. Represents the positive indicator in the positive indicator set The result after conversion; For the neutral indicator set, the following transformations are performed: ; Neutral indicator The average value of the time series corresponding to the time of collection, Neutral indicator The standard deviation of the time series corresponding to the time of collection, Represents a neutral indicator in a set of neutral indicators The result after conversion; For negative indicator sets, the following conversions are performed: ; Negative indicator The average value of the time series corresponding to the time of collection, Negative indicator The standard deviation of the time series corresponding to the time of collection, Represents a negative indicator in a negative indicator set The result after conversion.

[0012] In a preferred embodiment, the pre-trained fitness evaluation model is a convolutional neural network model.

[0013] In a preferred embodiment, a housing suitability index evaluation system integrating social perception spatiotemporal data includes: A dimension data acquisition module, used to respectively acquire relevant indicators of residential areas from multiple preset target dimensions; The feature extraction module is used to extract features from the relevant indicator acquisition process of each target dimension; The indicator acquisition evaluation module is used to evaluate the difficulty of indicator acquisition and contribution of relevant indicators of each dimension based on the feature extraction results; The indicator screening module is used to screen the relevant indicators of the target dimension according to the difficulty of indicator acquisition and the evaluation results of contribution role, and generate an evaluation indicator set; The classification module is used to summarize the evaluation indicator sets of all target dimensions and divide them into a positive indicator set, a neutral indicator set and a negative indicator set according to the preset classification criteria; A directivity conversion module is used to convert the directivity of the positive index set and the neutral index set respectively to generate a positive index vector, a neutral index vector and a negative index vector; The suitability evaluation module is used to take the positive index vector, the neutral index vector and the negative index vector as input data, input the pre-trained suitability evaluation model, and output the residential suitability as the evaluation result.

[0014] Technical effects and advantages of the present invention: The present invention combines multi-dimensional data, such as social perception, transportation convenience, medical resources, and environmental pollution. The interaction of these data is comprehensively evaluated through intelligent algorithms to ensure that the key factors affecting residential suitability in each dimension are effectively considered. This method can more accurately reflect the comprehensive suitability of residential areas, enabling urban planners to understand the living quality of residential areas from multiple perspectives, thereby making more scientific planning decisions.

[0015] By introducing real-time data sources, such as traffic flow, air quality, noise level, etc., the present invention can realize dynamic evaluation of residential suitability in various areas of the city. This provides strong support for urban planning. Urban planners can adjust planning schemes in a timely manner according to changes in actual conditions, optimize resource allocation, and achieve more flexible and dynamic urban management.

[0016] By adopting an intelligent suitability evaluation model, the present invention can effectively handle the complex relationship between multi-dimensional data and generate accurate suitability scores. Compared with traditional empirical decision-making, the present invention can provide data-driven, scientific suitability evaluation results, providing strong decision-making support for urban planners. By rationally planning residential areas, resource waste can be avoided and the foresight and sustainability of urban development can be improved.

[0017] By comprehensively considering factors closely related to the quality of life of residents, such as social perception, spatiotemporal behavior, etc., this invention can help urban planners better understand the actual needs of residents and formulate more targeted planning schemes for different resident groups. This can not only improve the suitability of residential areas, but also enhance the life satisfaction and happiness of residents and promote the harmonious development of the community. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings; Figure 1 This is a schematic diagram of the residential suitability index evaluation method integrating social perception spatiotemporal data in the present invention.

[0019] Figure 2 This is a schematic diagram of the housing suitability index evaluation system integrating social perception spatiotemporal data in the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Reference Figure 1 - Figure 2 The following embodiments are obtained: Example 1 The residential suitability index evaluation method integrating social perception spatiotemporal data includes the following steps: Relevant indicators of residential areas are obtained from multiple preset dimensions. Taking the preset dimensions as the target, feature extraction is performed in the process of obtaining relevant indicators of each target dimension. The difficulty of obtaining indicators and the contribution role are evaluated based on the feature extraction results; various data affecting residential suitability are collected from different angles or fields (called "dimensions"). For example, obtaining data from dimensions such as social perception, transportation, and medical care can cover different aspects of the residential environment and ensure that the evaluation of residential suitability is comprehensive and objective. In this context, a dimension is a specific field or angle used for analysis, which represents a set of factors that affect residential suitability in a field.

[0022] In this assessment process, "residential areas" can be areas of different attributes, each of which has its own unique characteristics and needs. The following is a detailed description and examples of different residential area attributes: Urban village housing: usually refers to old villages or areas with incomplete renovation located in the city. The living conditions in this area are often relatively simple, but because it is close to the city center, the rent is low, which attracts a large number of migrant workers or low-income groups. Urban village housing includes two types: private houses built by villagers in urban villages: mainly refers to private self-built houses in urban villages in the city center or surrounding areas. This type of housing is usually built by local villagers or small owners, and is mainly used for leasing; urban village collectively built small property rights housing: usually built on collective land, unable to obtain a complete property certificate, with a relatively low selling price, and mostly located in the urban-rural fringe. Examples of related indicators: the degree of completeness of infrastructure (such as the completeness of water supply, power supply, and drainage systems); sanitation conditions (such as garbage disposal and public health facilities); public security conditions (such as crime rate, police coverage); transportation convenience (such as the distance to the main transportation hubs in the city and commuting time). Suitability impact: The residential suitability of urban villages usually pays more attention to the cost of living, infrastructure and public security conditions, so these specific indicators are given a higher weight in the suitability assessment.

[0023] Commercial housing: a type of housing that is market-oriented, built by developers and sold to the market. The group that buys commercial housing usually has high requirements for living environment, quality of life and supporting facilities. Commercial housing includes ordinary commercial housing: housing sold in the ordinary market, with clear property rights, high prices, and mostly located in high-quality areas in the city; apartments: a type of collective housing, a form of living with complete living facilities but only occupying a part of the building. Generally divided into residential apartments and serviced apartments. Examples of relevant indicators: accessibility of educational resources (such as the number and distance of nearby high-quality schools); green environment (such as the area and distribution of green spaces and parks); commercial supporting facilities (such as the distribution of shopping malls, catering and entertainment); noise pollution level (such as the quietness of the surrounding environment). Suitability influence: The suitability of commercial housing is mainly affected by the living environment, educational resources and commercial supporting facilities. The target population pays attention to the quality of life and environmental comfort, so these dimensions will be highlighted in the suitability evaluation.

[0024] Different dimensions reflect different aspects of the residential environment, such as social perception dimension: residents' satisfaction, quality of life, community safety, etc. Spatiotemporal behavior dimension: the behavior patterns of residents in different time periods and spaces. Transportation convenience dimension: transportation infrastructure, commuting time, and public transportation convenience. Medical resources dimension: the accessibility and quality of medical services. Crisis response dimension: emergency response capabilities in emergencies. Environmental pollution dimension: air quality, noise pollution, water pollution, etc. The purpose of feature extraction is to extract the most representative information from the original data. By analyzing the relevant data under each dimension, the key features that can describe and measure residential suitability are extracted. These features are used for subsequent indicator acquisition evaluation and indicator screening. Content of feature extraction: For each target dimension, the extracted information may include: Acquisition difficulty information: the difficulty of obtaining the data. Computing cost information: the computing resources required to process the data. Linear correlation information: the correlation between the data and the final residential suitability.

[0025] According to the difficulty of obtaining and the contribution evaluation, the relevant indicators of the target dimension are screened to obtain the evaluation indicator set of the target dimension; the characteristic data of each dimension are evaluated to determine which indicators have a greater contribution to residential suitability, are less difficult to obtain, and have a lower calculation cost. Based on these criteria, indicators with higher value can be screened out. Evaluation method: Acquisition difficulty evaluation: Evaluate according to the difficulty of obtaining data. Computing cost evaluation: Evaluate according to the cost (time, resource consumption) of processing these data. Linear correlation evaluation: Use the Pearson correlation coefficient to evaluate the relationship between each indicator and residential suitability. The difficulty of indicator acquisition and contribution evaluation can screen out the most meaningful indicators in each dimension. This step is to remove irrelevant or less contributing indicators to improve the accuracy and efficiency of suitability evaluation. The selection of indicators is based on three types of criteria: Cost evaluation value: Remove indicators with too high cost. Difficulty evaluation value: Remove indicators with too high acquisition difficulty. Linear evaluation value: Retain indicators that are highly correlated with residential suitability.

[0026] The difficulty level reflects the resources, time, computational cost, and complexity required to obtain a certain indicator. The following aspects are mainly considered: Resource cost: the economic investment required to obtain this indicator, including data purchase costs, equipment installation costs, data processing costs, etc.

[0027] Time cost: The time required to collect and process the indicator. For example, some real-time data may be acquired quickly, but data involving historical data or complex processing may take longer.

[0028] Computational difficulty: The computational complexity of processing the indicator data, such as the difficulty of data preprocessing, cleaning, and feature extraction, and the demand for computing resources.

[0029] The higher the difficulty, the more difficult it is to obtain and process the indicator, and the higher the cost. Therefore, difficulty assessment can help determine whether an indicator is worth incorporating into the suitability evaluation model to avoid reducing the efficiency of the model due to excessive cost or complexity.

[0030] The contribution role measures the importance of a certain indicator in affecting residential suitability, that is, the degree of influence of the indicator on the suitability evaluation results. It mainly includes the following aspects: Influence: The strength of the correlation between the indicator and housing suitability. For example, indicators such as medical resources and air quality have a more direct impact on housing suitability and usually have a higher contribution.

[0031] Positive or negative indicator: Determine whether the indicator has a positive, negative or neutral impact on suitability to clarify its directional role in suitability evaluation. For example, greening rate as a positive indicator has a positive contribution to suitability, while noise pollution as a negative indicator will reduce suitability.

[0032] Uniqueness of indicators: Some indicators may have unique influence on a specific residential attribute. For example, environmental pollution has a greater impact on the suitability of high-end residential properties, but a smaller impact on private homes in urban villages. This type of uniqueness can increase the contribution of the indicator in a specific context. The greater the contribution, the more critical the indicator is in assessing residential suitability. Giving priority to indicators with high contribution can help improve the accuracy and scientificity of the model.

[0033] In the “Indicator Acquisition Assessment”, “Difficulty” and “Contribution” are used in combination to ensure that the assessment model can efficiently and accurately select key indicators that have a greater impact on the residential suitability evaluation.

[0034] Screening process: Prioritize indicators with large contribution and moderate difficulty. These indicators have a large contribution to the suitability results and will not significantly increase the cost of acquisition and calculation.

[0035] Optimize model efficiency: For indicators with too high difficulty, even if their contribution is large, they can be flexibly selected according to the actual situation. Reasonably balance indicators with high contribution and low difficulty to improve overall evaluation efficiency.

[0036] Improve assessment accuracy: Indicators with large contribution can more accurately reflect the actual situation of residential suitability, while indicators with moderate difficulty ensure the feasibility and economy of the assessment system.

[0037] The evaluation indicator sets of all target dimensions are summarized and divided into positive indicator sets, neutral indicator sets, and negative indicator sets according to the preset classification standards; the indicators of each dimension are classified to clarify which indicators have a positive impact on residential suitability, which are neutral evaluations, and which have a negative impact. The classification of positive, negative, and neutral indicators helps to perform different treatments when calculating the suitability index later. Classification standards: Positive indicators: The larger the value, the higher the residential suitability, such as greening rate and accessibility of medical resources. Neutral indicators: Indicators with little or neutral impact, such as the height of the residential floor may have a positive impact on some people (such as a better view on a higher floor), but may have a negative impact on others (such as high-rise fire escape is more difficult). Therefore, as a neutral indicator, it will not directly affect the suitability of the residence, but will be concerned by some people. Negative indicators: The larger the value, the lower the residential suitability, such as noise pollution and air pollution.

[0038] It should be noted that the basic indicator set includes general indicators for housing, that is, indicators that all housing have, and special indicators corresponding to housing types, that is, indicators that are unique to a certain type of housing. In practical applications, different types of housing, such as urban villages and commercial housing, will have their own special indicators in suitability assessment. These special indicators will be specially considered in the screening and classification of evaluation indicators, making the suitability assessment results of each type of housing more accurate and targeted. With the addition of these special indicators, housing with different attributes can be more comparable in suitability scores, ensuring that the comparison between houses with different attributes is fair and reasonable.

[0039] In the residential suitability evaluation method of the present invention, in order to more comprehensively and accurately reflect the suitability of various types of residential houses on the market and meet the different needs of the rental and sale markets, the residential areas can be divided more clearly and meticulously. The following is an example of specific classification based on the attributes of different types of residential houses: Private houses in urban villages: mainly refer to private self-built houses in urban villages in the city center or surrounding areas. Such houses are usually built by local villagers or small owners and are mainly rented out. The construction standards are low, and the infrastructure and property management may be imperfect, but due to the geographical advantages and relatively low rents, they attract a large number of migrant workers to rent.

[0040] The suitability evaluation focuses on: the degree of infrastructure perfection: the completeness of infrastructure such as water supply, power supply, and drainage; safety and public security: whether public security management and fire safety measures are in place; transportation convenience: the distance from major transportation hubs; living convenience: the distribution of surrounding daily living facilities (such as supermarkets, restaurants, etc.). Suitable groups: floating population, temporary residents, low-income tenants, etc.

[0041] Small-property houses built in urban villages: Small-property houses refer to houses built in urban villages or on collective land in rural areas without government approval and without complete property rights certificates. Small-property houses have attracted some buyers due to their low prices and flexible area selection, but due to unclear property rights and weak legal protection, transaction risks are high.

[0042] The suitability evaluation focuses on: property rights and legal risks: whether the property rights are clear and the assessment of legal risks; infrastructure conditions: the degree of perfection of water supply, drainage, public service facilities, etc.; environmental quality: including indicators such as greening, noise, and air quality; accessibility of educational resources: whether it is close to schools and meets family needs. Suitable groups: buyers who want to purchase low-cost properties but do not have high requirements for property rights.

[0043] Commercial housing: houses built by developers and sold to the market, with full property rights, and the target groups are mostly middle- and high-income families or investors. Commercial housing focuses on the overall living environment and provides relatively complete supporting facilities.

[0044] The suitability evaluation focuses on: supporting facilities for life: including the richness of supporting facilities such as education, medical care, shopping, and entertainment; living environment: such as greening rate, noise level, air quality, etc.; transportation conditions: the convenience of transportation to work areas, schools, and commercial areas; community property management: property management service level and community safety. Suitable groups: homebuyers, investors, and family homebuyers who pursue a high quality of life.

[0045] Apartment: A type of collective housing, a form of residence with complete living facilities but only occupying a part of the building. Generally divided into residential apartments and serviced apartments.

[0046] The suitability evaluation focuses on: work convenience: transportation convenience close to high-tech industrial parks or office centers; quality of life: including environmental facilities such as greenery, air quality, and community facilities; community services and management: such as property service levels and community safety management; education and medical resources: accessibility of high-quality education resources and medical facilities. Suitable groups: homebuyers, investors, and professional technicians who pursue a high quality of life.

[0047] This division can more clearly distinguish the properties of different types of residences, meet the different needs of the rental and sale markets, and ensure the comprehensiveness and accuracy of the suitability assessment: private houses in urban villages and small-property houses are more inclined to meet the needs of migrant populations and low-cost housing purchases, and the suitability assessment will pay more attention to aspects such as living costs, infrastructure, and public security. Commercial housing and apartments focus on living environment, living facilities, and community management to meet the needs of middle- and high-income groups for high-quality life. Through this division, the suitability evaluation system can conduct accurate and reasonable suitability assessments based on the characteristics of different types of residences, making the suitability of different types of residences comparable, which is convenient for users to choose appropriate residential areas according to their own needs. At the same time, for urban managers, it can also effectively improve the scientificity and accuracy of urban planning and management decisions.

[0048] By incorporating the specific indicators of different types of houses into the corresponding evaluation system, the residential suitability evaluation method of the present invention can make the suitability of houses with different attributes comparable. Since the evaluation indicator set for each type of house has been reasonably screened and classified, and the same processing method is used in the calculation of the suitability index, the suitability scores of houses with different attributes are consistent, which is convenient for direct comparison. For example, when the suitability indexes of urban villages and commercial houses are 0.6 and 0.8 respectively, this score difference is based on the suitability requirements of different residential types, which can provide a reliable reference for urban planners or home buyers, so as to have a clearer understanding of the actual living conditions of various types of houses.

[0049] The positive index set, neutral index set, and negative index set are converted to obtain positive index vectors, neutral index vectors, and negative index vectors respectively. Then, the positive index vectors, neutral index vectors, and negative index vectors are input as input data into the pre-trained fitness evaluation model. The output of the fitness evaluation model is residential suitability. The indicators of different categories are converted to ensure that they can make reasonable evaluations within different suitability ranges, so that these indicators can be used in a unified suitability calculation model. The converted indicators are integrated into corresponding vectors, which represent various influencing factors in residential suitability assessment. Each vector reflects the contribution of the relevant dimension to suitability. The positive index vector, neutral index vector, and negative index vector are input as input data into the pre-trained fitness evaluation model. The output of the fitness evaluation model is that residential suitability is the ultimate goal of the whole process, that is, all indicator data are comprehensively evaluated through the pre-trained fitness evaluation model to obtain a comprehensive score for residential suitability. The suitability evaluation model can be a convolutional neural network, which can extract complex nonlinear relationships from positive, neutral, and negative vectors and give an accurate suitability evaluation.

[0050] The multiple preset dimensions refer to the social perception dimension, the spatiotemporal behavior dimension, the transportation convenience dimension, the medical resources dimension, the crisis response dimension, and the environmental pollution dimension. The social perception dimension refers to measuring the quality of life and community environment of a residential area through residents' feedback and the social environment of the community. This type of dimension reflects the subjective feelings of residents such as life satisfaction, sense of security in the community, and social atmosphere. Related indicators: Resident satisfaction survey: obtain residents' overall satisfaction with the living environment through questionnaires or interviews. Community safety: measure the safety of the community based on data such as crime rate, public security situation, and frequency of police patrols. Social interactivity: evaluate the degree of social activity among residents based on factors such as the number of community activities, resident participation, and neighborhood relations. The social perception dimension reflects the soft indicators of community life, which usually affect residents' life satisfaction and overall happiness.

[0051] The spatiotemporal behavior dimension measures the behavior patterns of residents in space and time, analyzing their daily activities, travel time, and the distribution of people in residential areas. This dimension helps to understand the travel patterns and activity frequency of residents in residential areas. Related indicators: Crowd density: Monitor the peak flow of people in a specific area through the division of time and space. Travel time: Evaluate the daily commuting time and distance of residents, measure the degree of traffic congestion and time cost. Residential area activity: Monitor the activity of the community in different time periods through real-time data to reflect the busyness of the residential area. The spatiotemporal behavior dimension helps to evaluate the combination of movement and stillness in residential areas, and is suitable for evaluating the balance between work and life rhythms.

[0052] The transportation convenience dimension refers to the transportation convenience from residential areas to daily destinations such as urban areas, commercial areas, workplaces, schools, etc. This includes analysis of public transportation coverage, traffic congestion, commuting time, etc. Related indicators: Public transportation accessibility: such as the distance of buses and subway stations, the frequency and coverage of public transportation. Road congestion: Analyze the congestion of major roads through traffic flow monitoring and real-time traffic data. Commuting time: Measures the time required for residents in residential areas to travel to major work areas, commercial areas, schools, etc. Transportation infrastructure: including road conditions, completeness of traffic signs, traffic management efficiency, etc. Transportation convenience directly affects residents' travel efficiency, commuting time and convenience of life, and is a key factor in choosing a residential area.

[0053] The medical resources dimension measures the accessibility and quality of medical services in residential areas. This includes the distribution of hospitals, clinics, and pharmacies, as well as the response speed and quality of medical services. Related indicators: Distribution density of medical facilities: such as the distance and number of hospitals, clinics, and emergency centers. Accessibility of medical resources: Assess the time and transportation convenience of residents from residential areas to the nearest hospital or clinic. Quality of medical services: Based on the scale of medical institutions, service scope, advanced equipment, and professional level of medical staff. Emergency response time: In an emergency, the response speed from the residential area to the nearest hospital or emergency center. The medical resources dimension is related to the health protection and life safety of residents, and is an important dimension in the suitability evaluation, especially for the elderly and family residents.

[0054] The crisis response dimension assesses the response capabilities of residential areas when faced with natural disasters, social crises or emergencies. It includes emergency plans, resource allocation, facility safety, efficiency of rescue response, etc. Related indicators: Distribution of emergency shelters: The number and distribution of emergency shelters or emergency assembly points in and around residential areas. Crisis event response time: The time it takes for rescue personnel to arrive when events such as fires, earthquakes, and floods occur. Disaster risk situation: Based on historical data, determine whether the residential area is prone to natural disasters such as floods, earthquakes, and fires. Community emergency drills: The frequency and participation of emergency drills conducted in the community reflect the community's preparedness to face crises. The crisis response dimension directly affects the safety of the residence, especially when faced with emergencies. Efficient crisis response capabilities can provide residents with higher security.

[0055] The environmental pollution dimension evaluates environmental factors such as air quality, noise pollution, and water pollution in residential areas. Good environmental quality has an important impact on the health and living comfort of residents. Related indicators: Air Quality Index (AQI): Measures the impact of air quality on health based on the content of pollutants in the air (such as PM2.5, PM10, etc.). Noise pollution level: Assess the interference of noise sources (such as traffic noise, industrial noise, etc.) in residential areas to residents' lives. Water quality testing: Assess the degree of pollution by testing the water quality of tap water or rivers in residential areas. Soil pollution: For residential areas in suburbs or close to industrial areas, the degree of soil pollution is also an important indicator of environmental quality. The degree of environmental pollution directly affects the health and quality of life of residents, especially for long-term residents and residents with health problems, environmental quality is a crucial factor.

[0056] A data indicator database is preset, which contains multiple sets of basic indicator sets. Each basic indicator set corresponds to a target dimension, and the basic indicator set contains all relevant indicators of the corresponding target dimension. The data indicator database is a pre-designed and stored database used to store all data indicators required for evaluating residential suitability. This database is the core foundation of the residential suitability evaluation system, which contains various data under multiple dimensions. The data in the database is divided by dimension, and each dimension corresponds to a basic indicator set. Each basic indicator set contains all data indicators related to residential suitability evaluation under that dimension.

[0057] The basic indicator set is a list of predefined indicators for each dimension, which includes all the indicators required to evaluate the dimension. Each basic indicator set is associated with a target dimension and contains all the factors that affect residential suitability under that dimension. The data in the basic indicator set may include raw data (such as real-time data, historical data) and derived indicators (such as complex indicators calculated by algorithms). The data indicator database associates each basic indicator set with a specific target dimension. The target dimension is the different angles or fields for evaluating residential suitability, such as social perception, convenient transportation, medical resources, etc.

[0058] Examples of data indicator types in the basic indicator set: Quantitative indicators: such as the number of hospitals, distance to transportation stations, air quality index (AQI), etc., these indicators have a clear numerical range. Qualitative indicators: such as community satisfaction surveys, sense of security index, etc., these indicators are mainly based on subjective evaluation. Derived indicators: indicators generated by calculation or modeling of quantitative data, such as traffic congestion index, environmental risk assessment value, etc. Derived indicators are implemented through calculation models in the prior art and will not be elaborated on here. Since they involve computing power, they will need to be screened later, which can save computing power and shorten the time spent on the implementation of the present invention.

[0059] Taking the preset dimension as the target, when extracting features for the relevant indicator acquisition process of each target dimension, the feature extraction includes acquisition difficulty information, computing cost information and linear related information. Evaluating the difficulty and contribution of indicator acquisition based on feature extraction results means: The difficulty evaluation value is generated based on the difficulty information obtained. The calculation formula is: ; is the difficulty evaluation value of the ith indicator, To obtain the time required for the i-th indicator, The economic cost of obtaining the i-th indicator is: The number of steps required to obtain the i-th indicator, is the availability coefficient of the ith indicator, , , These are preset difficulty impact coefficients, corresponding to the weights of time, cost, and steps, respectively, indicating the relative importance of each factor in the difficulty assessment; time cost Expressed in exponential form, it means that as the acquisition time increases, the acquisition difficulty increases exponentially. Longer time will make the difficulty rise rapidly. Time is a core difficulty factor in acquiring indicators, and the exponential form reflects the sensitivity of time delay to the overall difficulty. Economic cost The square root form means that the effect of cost on difficulty is nonlinear, higher cost increases difficulty, but not as fast as time exponentially. The constant 1 is added to avoid the situation where the difficulty is 0 when the cost is very small or zero. Economic cost has some effect on difficulty, but it does not change as drastically as time factors, and the square root form makes its contribution to difficulty smoother. Formula part , the ratio of the number of steps to the availability coefficient is used to measure the complexity of the acquisition steps. If the number of steps is large and the availability coefficient is low, the difficulty value will increase significantly. The more steps and the worse the availability of the indicator, the more difficult it is to obtain. This part reflects the contribution of acquisition complexity to the overall difficulty through linear processing. The larger the difficulty assessment value, the more difficult it is to obtain the indicator. This indicator requires more time, economic cost and operation steps to be obtained, and the acquisition process may be more complex and challenging. The longer it takes to obtain, the greater the difficulty. The higher the economic cost of obtaining this indicator, the greater the difficulty, but its impact is relatively smooth. The more steps and the lower the availability coefficient, the greater the difficulty.

[0060] Availability factor The calculation is performed by weighted summing the proportion of the area covered during the data source collection process corresponding to the i-th indicator to the entire residential area, i.e., the coverage area proportion A, the data source collection frequency B, and the data source completeness C; ; , , are all preset weight coefficients. The coverage area ratio A indicates the proportion of the area covered by the data source of the ith indicator to the entire residential area. The larger the coverage ratio, the more areas the data source can provide data in, and the higher the data availability. If a data source covers more residential areas, its availability is higher, which means that the contribution of this data source is more reliable when evaluating suitability. For example, if the noise monitoring equipment covers the entire community instead of only a small area, the noise data will be more reliable and representative. The value of the coverage area ratio is usually between [0, 1], indicating the degree from partial coverage to full coverage. The data source collection frequency indicates the data collection frequency of the ith indicator. For example, real-time collection, daily collection, monthly collection, etc. The higher the collection frequency, the more timely the data update and the more accurately it can reflect the current residential status. Frequent data collection means that the system can obtain more real-time data and can reflect changes in the residential environment more timely. For example, real-time air quality monitoring can better reflect the actual situation than data collected once a month. The value of the data source collection frequency B can be set according to the frequency of data collection. For example, real-time data can be set to 1, and lower frequency data can be set to 0.1 or 0.2. The completeness C of the data source indicates the completeness of the data source of the ith indicator, reflecting whether the data source has any missing data, whether the records are complete, and other issues. The higher the completeness, the better the data quality of the data source and the higher the availability. High-completeness data means that they can more accurately reflect residential conditions. If a data source has more missing data or noise interference, its availability will be reduced. Data sources with high completeness provide more reliable results and more accurate data evaluation. The value of the completeness C of the data source is usually between [0, 1], indicating the degree of data integrity from partial missing to complete.

[0061] The cost evaluation value is generated based on the computing power cost information. The calculation formula is: ; is the cost evaluation value of the ith indicator, is the time required to calculate the i-th indicator, is the amount of data required for the i-th indicator, is the efficiency of parallel computing for the i-th indicator, , , These are preset cost impact coefficients, which are used to control the weights of time, data volume, and computing efficiency in the total cost. The computing time indicates the time required to process or calculate the indicator. The importance of time cost is reflected in the fact that if the computing time is long, the overall computing cost will be significantly increased. The formula part is , the time cost in the formula increases in a square form, indicating that the impact of time on computing power cost is nonlinear. As the computing time increases, the cost shows an accelerating growth trend. Used to adjust the contribution of computing time to the overall computing cost. The longer the computing time, the greater the computing cost assessment value, especially when the time increases exponentially, the cost will increase more significantly. Data volume refers to the amount of data required to process this indicator, usually referring to the size of data storage, the number of data packets transmitted, etc. The larger the data volume, the more storage space and bandwidth are required for processing.

[0062] The formula part is , the effect of data volume on computing power cost is processed by logarithmic function, which shows that the effect of increasing data volume on cost is gradually decreasing. In other words, the larger the data volume, the more gradual the impact of its increase on cost. Adding 1 is to avoid the problem that the logarithm cannot be calculated when the data volume is zero. Coefficient Used to adjust the weight of data volume in computing power cost. The increase of data volume will increase computing power cost, but its growth impact is more gradual than time cost. It is suitable for scenarios with large data volume but high computing efficiency.

[0063] Computational efficiency refers to the efficiency of parallel computing. The higher the value, the higher the computing efficiency and the shorter the processing time. Parallel computing can share the computing burden and thus reduce the computing cost. The formula is , the impact of computing efficiency on computing power cost is inversely proportional. In other words, the higher the computing efficiency, the lower the computing power cost. The weight of control efficiency in computing power cost. Through inverse form processing, the formula can reflect the effect of efficiency improvement on reducing overall computing power cost. Computational efficiency is an important factor in reducing computing power cost. The higher the efficiency, the lower the computing power cost assessment value. The computing power cost assessment value is a numerical value that measures the acquisition and calculation cost of a certain indicator. The larger the value, the more computing resources are required to obtain the indicator, and the more complex the calculation process for processing the indicator. The longer the time, the higher the computing power cost, especially when the time increases exponentially, the computing power cost will rise rapidly. The larger the amount of data, the higher the computing power cost, but its impact gradually tends to be flat. The higher the efficiency, the lower the computing power cost, so the optimization of parallel processing can effectively reduce the computing power cost.

[0064] Based on the linear correlation information, a linear evaluation value is generated, and the Pearson correlation coefficient is used to evaluate the correlation between each indicator and residential suitability to obtain the correlation coefficient. , the linear evaluation value is marked as And the value and correlation coefficient The absolute value of is equal to that of the other two indicators. The linear evaluation value indicates the degree of influence of the indicator in the suitability assessment after evaluating the linear relationship between an indicator and residential suitability through the Pearson correlation coefficient. The linear evaluation value only takes the absolute value of the correlation coefficient, which means that the focus is on the strength of the correlation, not whether it is positive or negative. In other words, no matter whether an indicator has a positive or negative impact on residential suitability, as long as the impact is strong, the indicator has a higher linear evaluation value. The larger the linear evaluation value, the stronger the linear relationship between the indicator and residential suitability, indicating that the indicator has a greater impact on residential suitability. Conversely, the smaller the linear evaluation value, the weaker the relationship between the indicator and residential suitability and the smaller the influence.

[0065] According to the difficulty of obtaining indicators and the evaluation results of contribution, the relevant indicators of the target dimension are screened, and the evaluation indicator set of the target dimension is: Compare the cost evaluation value of each indicator of the target dimension with the preset threshold value 1, eliminate indicators that are greater than or equal to the preset threshold value 1, and obtain indicator set 1; indicators with higher cost evaluation values ​​mean that the resource consumption (time, computing power, data volume, etc.) to obtain these indicators is greater, so if the cost is too high, the actual application value of these indicators will be relatively low. Eliminating indicators with too high costs can ensure that the indicators used for evaluation are economically feasible and reduce unnecessary waste of resources during the evaluation process.

[0066] Compare the difficulty evaluation value of each indicator of the target dimension with the preset threshold value 2, eliminate indicators that are greater than or equal to the preset threshold value 2, and obtain indicator set 2; the difficulty evaluation value reflects the complexity of obtaining a certain indicator, including acquisition time, economic cost, data acquisition difficulty, etc. Eliminate indicators with too high difficulty to ensure that the system only retains those indicators with relatively low acquisition difficulty, reduce operational complexity, and improve evaluation efficiency.

[0067] Compare the linear evaluation value of each indicator of the target dimension with the preset threshold value three, eliminate the indicators that are less than the preset threshold value three, and obtain the indicator set three; the linear evaluation value measures the correlation between a certain indicator and residential suitability, reflecting the explanatory power of the indicator for residential suitability. Eliminate indicators with weak linear correlation to ensure that the remaining indicators have strong explanatory power for suitability assessment, thereby improving the scientificity and accuracy of the assessment.

[0068] The intersection of indicator set 1, indicator set 2 and indicator set 3 is obtained as the evaluation indicator set for this target dimension. The intersection operation ensures that the indicators retained in the end have good performance in three aspects (cost, difficulty, and relevance). By comprehensively considering different screening conditions, a set of indicators that are economically feasible, easy to obtain, and have a significant contribution to the accuracy of the evaluation are finally selected. These indicators represent the best choice, which can ensure that when evaluating residential suitability, it is scientific and does not waste too many resources.

[0069] Directivity switching operation refers to: For the forward indicator set, the following transformations are performed: ; A positive indicator The standard deviation of the time series corresponding to the time of collection, A positive indicator The average value of the time series corresponding to the time of collection, Indicates the offset set to prevent meaningless calculations. Represents the positive indicator in the positive indicator set The result after conversion; positive indicators refer to indicators with larger values, which have a more positive impact on residential suitability. For example, greening rate, accessibility of medical resources, etc. For these indicators, the larger the value, the higher the suitability. Using logarithmic transformation can narrow the range of indicators with larger values ​​to avoid excessive values ​​dominating the final results, and smooth the data to avoid excessive fluctuations. The standard deviation term reflects the volatility of the data. The greater the volatility, the more unstable the indicator is, and the suitability will be adversely affected. Therefore, the term minus the standard deviation is used to penalize indicators with large volatility. The larger the value of the converted positive indicator, the greater the positive impact of the indicator on residential suitability; at the same time, if the standard deviation is large, the positive impact of the indicator will be reduced.

[0070] For the neutral indicator set, the following transformations are performed: ; Neutral indicator The average value of the time series corresponding to the time of collection, Neutral indicator The standard deviation of the time series corresponding to the time of collection, Represents a neutral indicator in a set of neutral indicators The result after conversion; it should be noted that when the log function in the present invention does not have a clear base, it should be interpreted as a natural logarithm (base e). A neutral indicator refers to an indicator whose impact on suitability is neither completely positive nor completely negative. The value of such an indicator has little effect on suitability, or its impact depends on the specific scenario. For example, residential height, number of parking spaces, etc., some residents may think these factors are important, while other residents do not care. Logarithmic transformation of neutral indicators and normalization according to their standard deviation means that we neither over-amplify the impact of these indicators nor ignore their volatility. Standardization ensures that its impact on suitability assessment is relatively neutral. Neutral indicators can maintain their neutrality after conversion, but their influence will still be adjusted according to their volatility. The larger the value, the more positive the impact of the neutral indicator on suitability.

[0071] For negative indicator sets, the following conversions are performed: ; Negative indicator The average value of the time series corresponding to the time of collection, Negative indicator The standard deviation of the time series corresponding to the time of collection, Represents a negative indicator in a negative indicator set The result after conversion. Negative indicators refer to indicators whose larger values ​​have a more negative impact on residential suitability. For example, noise pollution, air pollution, etc. These indicators have a negative impact on suitability, and the smaller the value, the better. Use the inverse form to ensure that the larger the value of the negative indicator, the smaller the value after conversion, which has a negative impact on suitability. If the indicator value is small, the converted value is large, and the negative impact on suitability is smaller. Like positive indicators, the standard deviation reflects the volatility of the data. The greater the volatility, the more unstable the negative indicator is. Therefore, it is necessary to punish negative indicators with high volatility to reduce their contribution to suitability. After the negative indicator is converted, the larger the value, the smaller the negative impact of the indicator. As the original value increases, its negative impact on suitability increases.

[0072] Different types of indicators affect suitability in different ways, and these effects cannot be reasonably reflected by directly using raw data. Through these directional transformation operations, the role of positive, negative and neutral indicators in suitability assessment can be more effectively reflected: positive indicators are logarithmically transformed to avoid excessive values ​​affecting the overall assessment, and the standard deviation is subtracted to penalize volatility, ensuring that only stable and positive indicators can positively affect suitability. Neutral indicators are standardized to maintain neutral influence and avoid excessive impact on the assessment results. Negative indicators are inversely processed, and the larger the value, the more unfavorable it is to suitability, ensuring that the negative impact of negative indicators on suitability can be reasonably reflected in the calculation.

[0073] The pre-trained fitness evaluation model is a convolutional neural network model. The convolutional neural network model (CNN) was selected as the pre-trained fitness evaluation model because it performs well in processing multi-dimensional and complex data, and is particularly suitable for analyzing data with spatial structure or local correlation. For residential suitability assessment, data from multiple dimensions are involved, such as social perception, spatiotemporal behavior, convenient transportation, medical resources, crisis response, and environmental pollution. There may be complex nonlinear relationships within each dimension and between different dimensions. Convolutional neural networks can effectively capture these relationships through their structure, thereby providing more accurate prediction results for suitability evaluation.

[0074] In the process of suitability evaluation, the use of convolutional neural networks can be divided into the following steps: Input data: Multi-dimensional indicators related to residential areas, such as social perception, spatiotemporal behavior, convenient transportation, medical resources, crisis response, and environmental pollution, are pre-processed and used as input data into the CNN model. These input data may include various positive, neutral, and negative indicators, which have been standardized and uniformly formatted through the previous directional conversion operation, suitable for direct use in model processing. The data of each dimension can be regarded as a layer of feature maps. These data are used as the input layer of CNN. The input data can be a two-dimensional or three-dimensional structure, such as a multi-dimensional data matrix corresponding to each residential area.

[0075] Convolutional layer extracts features: After the input data enters CNN, it first passes through multiple convolutional layers. Each convolutional layer extracts local features through convolution operations and activation functions, capturing patterns and correlations of data of different dimensions layer by layer. For example: the first layer of convolution may extract features reflecting community safety and resident satisfaction from the social perception dimension; the second layer of convolution may extract features of commuting time and traffic congestion from the traffic dimension; as the convolutional layer goes deeper, the model will combine the data features of each dimension to identify some high-level patterns, such as the potential correlation between medical resources and convenient transportation, or the interactive effects of social perception and environmental pollution. The role of the convolutional layer is to gradually transform high-dimensional input data into a series of meaningful feature representations, so that complex data structures can be effectively represented by a small number of features.

[0076] Pooling layer and dimensionality reduction: After each convolutional layer, a pooling layer (such as a maximum pooling layer or an average pooling layer) is usually introduced to reduce the dimensionality and size of the feature map. This operation can remove redundant information, retain the most important features, reduce the amount of model computation, and prevent the model from overfitting.

[0077] Fully connected layer for classification or regression: After multiple layers of convolution and pooling, the final feature map is flattened and input to the fully connected layer. The fully connected layer performs classification or regression prediction based on the features extracted by the convolution layer. In the residential suitability assessment, the final output is a suitability score. The suitability score represents the comprehensive suitability of a residential area. This score can reflect the result of comprehensive consideration of data in various dimensions and is used to measure the residential suitability of the area.

[0078] Output: The final output of the CNN model is a suitability score for a residential area. This score can be a value between 0 and 1, where 1 is the most suitable place to live and 0 is completely unsuitable. The output suitability score can be directly used as a core indicator in the evaluation report to judge the living quality of a residential area.

[0079] Application scenarios: Residential developers, policymakers or individual homebuyers can choose the most suitable area based on this suitability score, or take corresponding measures to improve the living environment in certain areas. The suitability score can help homebuyers make wise home purchase decisions based on the scoring results of different areas. For example, a suitability score of 0.85 in a certain area means that the area is very suitable for living and has high-quality transportation, medical and community environment. The government or developers can use the suitability score to evaluate and improve existing residential areas. Low suitability areas may need to increase public facilities, improve environmental pollution or optimize traffic planning. Real estate investors can evaluate the potential value of an area based on the suitability score. High suitability areas usually mean higher market demand and future value-added space.

[0080] Example 2 The residential suitability index evaluation system integrating social perception spatiotemporal data includes: A dimension data acquisition module, used to respectively acquire relevant indicators of residential areas from multiple preset target dimensions; The feature extraction module is used to extract features from the relevant indicator acquisition process of each target dimension; The indicator acquisition evaluation module is used to evaluate the difficulty of indicator acquisition and contribution of relevant indicators of each dimension based on the feature extraction results; The indicator screening module is used to screen the relevant indicators of the target dimension according to the difficulty of indicator acquisition and the evaluation results of contribution role, and generate an evaluation indicator set; The classification module is used to summarize the evaluation indicator sets of all target dimensions and divide them into a positive indicator set, a neutral indicator set and a negative indicator set according to the preset classification criteria; A directivity conversion module is used to convert the directivity of the positive index set and the neutral index set respectively to generate a positive index vector, a neutral index vector and a negative index vector; The suitability evaluation module is used to take the positive index vector, the neutral index vector and the negative index vector as input data, input the pre-trained suitability evaluation model, and output the residential suitability as the evaluation result.

[0081] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0082] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0083] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0084] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0085] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A residential suitability index evaluation method integrating social perception spatiotemporal data, characterized in that: The following steps are involved: Relevant indicators of residential areas are obtained from multiple preset dimensions respectively. Taking the preset dimensions as the target, the acquisition process of relevant indicators of each target dimension is subjected to feature extraction, and the difficulty of indicator acquisition and contribution are evaluated based on the feature extraction results; Screening the relevant indicators of the target dimension according to the difficulty of acquisition and the evaluation results of contribution role to obtain a set of evaluation indicators of the target dimension; Summarize the evaluation indicator sets of all target dimensions and divide them into positive indicator sets, neutral indicator sets, and negative indicator sets according to the preset classification standards; The positive indicator set, the neutral indicator set, and the negative indicator set are respectively subjected to directional conversion operations to obtain a positive index vector, a neutral index vector, and a negative index vector, and then the positive index vector, the neutral index vector, and the negative index vector are input together as input data into a pre-trained fitness evaluation model. The output result of the fitness evaluation model is the residential suitability.

2. The residential suitability index evaluation method integrating social perception spatiotemporal data according to claim 1 is characterized in that: The multiple preset dimensions refer to the social perception dimension, the spatiotemporal behavior dimension, the transportation convenience dimension, the medical resources dimension, the crisis response dimension, and the environmental pollution dimension.

3. The residential suitability index evaluation method integrating social perception spatiotemporal data according to claim 2 is characterized in that: A data indicator database is preset, which includes multiple groups of basic indicator sets. Each basic indicator set corresponds to a target dimension, and the basic indicator set includes all relevant indicators of the corresponding target dimension.

4. The residential suitability index evaluation method integrating social perception spatiotemporal data according to claim 3 is characterized in that: Taking the preset dimension as the target, when the process of acquiring relevant indicators of each target dimension is subjected to feature extraction, the feature extraction includes acquisition difficulty information, computing power cost information and linear related information.

5. The residential suitability index evaluation method integrating social perception spatiotemporal data according to claim 4 is characterized in that: The evaluation of the difficulty of obtaining indicators and their contribution based on feature extraction results refers to: The difficulty evaluation value is generated based on the difficulty information obtained. The calculation formula is: ; is the difficulty evaluation value of the ith indicator, To obtain the time required for the i-th indicator, The economic cost of obtaining the i-th indicator is: The number of steps required to obtain the i-th indicator, is the availability coefficient of the ith indicator, , , These are all preset difficulty influence coefficients; Availability factor It is calculated by weighted summing the proportion of the area covered by the data source collection process corresponding to the i-th indicator to the entire residential area, the data source collection frequency, and the completeness of the data source; The cost evaluation value is generated based on the computing power cost information. The calculation formula is: ; is the cost evaluation value of the ith indicator, is the time required to calculate the i-th indicator, is the amount of data required for the i-th indicator, is the efficiency of parallel computing for the i-th indicator, , , All are preset cost impact coefficients; Based on the linear correlation information, a linear evaluation value is generated, and the Pearson correlation coefficient is used to evaluate the correlation between each indicator and residential suitability to obtain the correlation coefficient. , the linear evaluation value is marked as And the value and correlation coefficient The absolute values ​​are equal.

6. The residential suitability index evaluation method integrating social perception spatiotemporal data according to claim 5 is characterized in that: According to the difficulty of obtaining indicators and the evaluation of contribution, the relevant indicators of the target dimension are screened, and the evaluation indicator set of the target dimension is: Compare the cost evaluation value of each indicator of the target dimension with a preset threshold value 1, eliminate indicators that are greater than or equal to the preset threshold value 1, and obtain an indicator set 1; Compare the difficulty evaluation value of each indicator of the target dimension with the preset threshold value 2, eliminate indicators that are greater than or equal to the preset threshold value 2, and obtain indicator set 2; Compare the linear evaluation value of each indicator of the target dimension with the preset threshold value three, eliminate the indicators that are less than the preset threshold value three, and obtain the indicator set three; The intersection of indicator set 1, indicator set 2 and indicator set 3 is obtained as the evaluation indicator set of the target dimension.

7. The residential suitability index evaluation method integrating social perception spatiotemporal data according to claim 6 is characterized in that: Directivity switching operation refers to: For the forward indicator set, the following transformations are performed: ; A positive indicator The standard deviation of the time series corresponding to the time of collection, A positive indicator The average value of the time series corresponding to the time of collection, Indicates the offset set to prevent meaningless calculations. Represents the positive indicator in the positive indicator set The result after conversion; For the neutral indicator set, the following transformations are performed: ; Neutral indicator The average value of the time series corresponding to the time of collection, Neutral indicator The standard deviation of the time series corresponding to the time of collection, Represents a neutral indicator in a set of neutral indicators The result after conversion; For negative indicator sets, the following conversions are performed: ; Negative indicator The average value of the time series corresponding to the time of collection, Negative indicator The standard deviation of the time series corresponding to the time of collection, Represents a negative indicator in a negative indicator set The result after conversion.

8. The residential suitability index evaluation method integrating social perception spatiotemporal data according to claim 7 is characterized in that: The pre-trained fitness evaluation model is a convolutional neural network model.

9. A residential suitability index evaluation system integrating social perception spatiotemporal data, which is implemented based on the residential suitability index evaluation method integrating social perception spatiotemporal data according to any one of claims 1 to 8, characterized in that: include: A dimension data acquisition module, used to respectively acquire relevant indicators of residential areas from multiple preset target dimensions; The feature extraction module is used to extract features from the relevant indicator acquisition process of each target dimension; The indicator acquisition evaluation module is used to evaluate the difficulty of indicator acquisition and contribution of relevant indicators of each dimension based on the feature extraction results; The indicator screening module is used to screen the relevant indicators of the target dimension according to the difficulty of indicator acquisition and the evaluation results of contribution role, and generate an evaluation indicator set; The classification module is used to summarize the evaluation indicator sets of all target dimensions and divide them into a positive indicator set, a neutral indicator set and a negative indicator set according to the preset classification criteria; A directivity conversion module is used to convert the directivity of the positive index set and the neutral index set respectively to generate a positive index vector, a neutral index vector and a negative index vector; The suitability evaluation module is used to take the positive index vector, the neutral index vector and the negative index vector as input data, input the pre-trained suitability evaluation model, and output the residential suitability as the evaluation result.

Images