Method and system for determining permanent residence of floating population based on agent graph and medium

Abstract

The application provides a method and system for determining the permanent residence of floating population based on an intelligent agent graph and a medium, and belongs to the technical field of floating population data analysis and government data graph mining. The method constructs a standardized dataset by collecting multi-source spatio-temporal and social correlation data of floating population, builds an intelligent agent graph with dynamic weight correlation edges, extracts three types of permanent residence determination features based on graph subgraph clustering and time series weight evolution, combines region-specific and population-specific sample training of adaptive weight determination model and differentiated threshold, and completes the permanent residence determination by fusing feature weighting calculation and comprehensive score. The application relies on the intelligent agent graph to depict the complete territorial behavior and social binding relationship of the floating population, discards the fixed determination threshold, can adapt to the flow differences of different regional populations, weakens the interference of short-term temporary behavior, improves the accuracy and scene adaptability of the permanent residence determination, and provides reliable digital determination basis for the normal governance of urban population.

Description

Technical Field

[0001] This application relates to the fields of migrant population data analysis and government big data graph mining technology, specifically to a method, system, and medium for determining the permanent residency of migrant population based on intelligent agent graphs. Background Technology

[0002] Current methods for identifying the permanent residence status of the floating population largely rely on single trajectory duration thresholds. This results in limited data dimensions, a lack of social relationship modeling, and susceptibility to interference from temporary travel and short-term stays. Patent CN112001829B, "A Method for Determining Population Distribution Based on Mobile Phone Signaling Data," identifies permanent residents solely based on base station dwell time statistics. It uses only single signaling spatiotemporal data, failing to incorporate socially relevant information such as employment, social interaction, and public service connections. It lacks a logic for building a person-place relationship graph and relies solely on fixed duration thresholds to differentiate population types, making it unsuitable for varying regional migration patterns. Patent CN115840856A, "A Method and Device for Identifying the Population in a Jurisdiction Based on Spatiotemporal Big Data," only collects basic housing and personnel registration data, failing to explore behavioral evolution patterns and lacking dynamic weight graph construction and adaptive weight training mechanisms. It cannot quantify the stability of long-term resident behavior, and temporary visit data easily leads to distorted judgments. These methods suffer from technical deficiencies such as limited dimensions, poor scenario adaptability, and weak anti-interference capabilities.

[0003] There is an urgent need for a solution to determine the permanent residency of the floating population that integrates multi-source behavioral and social data, relies on intelligent agent graphs to quantify relationships, and supports adaptive differentiated determination. This solution would address the problems of existing technologies, such as single dimensions, static modeling, fixed thresholds, and distorted judgments, and would be suitable for the needs of refined and high-precision population identification in various scenarios. Summary of the Invention

[0004] The purpose of this application is to provide a method, system, and medium for determining the permanent residency of the floating population based on an intelligent agent graph. This method collects multi-source spatiotemporal behavioral and social relational data of the floating population, performs standardized preprocessing, constructs an intelligent agent graph with dynamically weighted relational edges using the floating population, local locations, and social service institutions as entity nodes, and mines three core determination features—spatiotemporal residency, behavioral inertia, and social solidification—based on the graph topology and temporal data mining. It trains an adaptive weighted determination model and a differentiated threshold system using region- and population-specific labeled samples, weights the extracted core features to obtain a comprehensive permanent residency score, matches the corresponding threshold, and outputs the determination result. Simultaneously, it includes incremental graph updates and abnormal behavior noise reduction and correction optimization paths, fully realizing multi-dimensional, adaptive, and interference-resistant automated determination of the permanent residency of the floating population.

[0005] This application provides a method for determining the permanent residency of the floating population based on agent graphs, including the following steps: Collect multi-source spatiotemporal behavioral data and social correlation data of the target migrant population within a preset statistical period, and clean, deduplicate, and spatiotemporally align the data according to preset data preprocessing rules to obtain a standardized migrant population behavior dataset. Collect entity node information of migrant population, local locations, and social service organizations; mine the relationships between nodes based on standardized migrant population behavior datasets; assign edge weights based on behavior frequency, duration, and stability parameters to obtain a migrant population intelligent agent graph with dynamically weighted associated edges. The system collects the subgraph structure, edge weight time series data, and node association data of the intelligent agent graph. Based on subgraph clustering, weight evolution analysis, and association degree calculation, it obtains three core characteristics for determining permanent residency: spatiotemporal residence, behavioral inertia, and social solidification. Collect historical resident-labeled sample data of the floating population, train the model based on the behavioral characteristics corresponding to regional environmental differences and population attribute differences, and obtain a judgment weight model and a differentiated threshold system that can adaptively adjust feature weights. Based on the three core characteristics of permanent residency—spatiotemporal residence, behavioral inertia, and social solidification—a weighted calculation is performed using the adaptively adjustable characteristic weight model to obtain a comprehensive score for the permanent residency of the floating population. Based on the comprehensive score of the permanent residency of the floating population, the permanent residency determination result and corresponding confidence level of the target floating population are obtained by comparison and identification through the differentiated threshold system.

[0006] In the method for determining the permanent residency of the floating population based on intelligent agent graphs described in this application, the collection of multi-source spatiotemporal behavioral data and social association data of the target floating population within a preset statistical period, followed by data cleaning, deduplication, and spatiotemporal alignment according to preset data preprocessing rules, yields a standardized floating population behavior dataset. Specifically: Collect raw spatiotemporal data on the local residence, travel trajectory, and local consumption behavior of the target migrant population within a preset statistical period; Collect raw data on social connections of the target migrant population, including their local employment, public service registration, and local social connections. The original spatiotemporal behavior data and social relational data are processed according to preset data compliance verification and duplicate data removal rules to obtain cleaned and effective data. The cleaned and valid data are time-series aligned based on a unified spatiotemporal coordinate benchmark to obtain a standardized dataset of migrant population behavior with regularized dimensions.

[0007] In the method for determining the permanent residency of the floating population based on the agent graph described in this application, the steps of collecting entity node information of the floating population, local locations, and social service organizations, mining the relationships between nodes based on a standardized floating population behavior dataset, and assigning edge weights based on behavior frequency, duration, and stability parameters to obtain a floating population agent graph with dynamically weighted associated edges are as follows: Collect basic attribute data of entity nodes corresponding to individuals of the floating population, local activity venues, and local social service organizations; Based on the standardized migrant population behavior dataset, the association mapping relationship between each entity node is mined. The weights of the associated edges of nodes are quantified based on the frequency of behavior occurrence, cumulative dwell time, and temporal fluctuation stability parameters. By binding each entity node with dynamically assigned associated edge weights, a complete intelligent agent graph of the floating population is obtained.

[0008] In the method for determining the permanent residency of the floating population based on agent graphs described in this application, the collection of subgraph structure, edge weight time-series data, and node association data of the agent graph, and the determination of three core permanent residency characteristics—spatiotemporal residence, behavioral inertia, and social solidification—are obtained based on subgraph clustering, weight evolution analysis, and association degree calculation methods. Specifically: Collect the subgraph topology and node association density data of the intelligent agent graph of the floating population; Collect time-series weight evolution data of the associated edges of the mobile population intelligent agent graph, and record the weight fluctuation and change status within the period; Based on subgraph clustering, temporal evolution analysis, and node correlation calculation methods, the multidimensional features of the graph are quantified respectively. Based on the quantification results of the features, three core judgment features were obtained: spatiotemporal dwell time, behavioral inertia, and social solidification.

[0009] In the method for determining the permanent residency of the floating population based on intelligent agent graphs described in this application, the step of collecting historical floating population permanent residency labeled sample data, training a model based on behavioral characteristic patterns corresponding to regional environmental differences and population attribute differences, and obtaining a determination weight model and a differentiated threshold system that can adaptively adjust feature weights, specifically involves: Collect manually labeled historical permanent residence data of the floating population in different regions and functional scenarios; Collect behavioral characteristic distribution data of migrant population samples with different age structures and employment attributes; The initial model parameters are iteratively trained based on the differences in behavioral characteristics corresponding to regional environment and population attributes. Based on the converged model parameters, an adaptive weight adjustment model and a differentiated threshold system for matching scenarios are obtained.

[0010] In the method for determining the permanent residency of the floating population based on intelligent agent graphs described in this application, the step of performing a weighted calculation based on the three core characteristics of permanent residency—spatiotemporal residence, behavioral inertia, and social solidification—through a determination weight model with adaptively adjustable feature weights to obtain a comprehensive score for the permanent residency of the floating population is as follows: Collect quantitative values ​​of the three core judgment features: spatiotemporal dwell time, behavioral inertia, and social solidification; Collect the real-time weight configuration parameters of the trained adaptive decision weight model; The three core features are multidimensionally weighted and fused according to the real-time weight configuration parameters of the adaptive judgment weight model to obtain the initial score data; The initial scoring data is processed according to a preset normalization correction rule to obtain a standardized comprehensive score for the permanent residence of the floating population.

[0011] Secondly, this application provides a system for determining the permanent residency of a floating population based on an intelligent agent graph, including a memory and a processor. The memory includes a program for determining the permanent residency of a floating population based on an intelligent agent graph. When the processor executes the program for determining the permanent residency of a floating population based on an intelligent agent graph, it performs the following steps: Collect multi-source spatiotemporal behavioral data and social correlation data of the target migrant population within a preset statistical period, and clean, deduplicate, and spatiotemporally align the data according to preset data preprocessing rules to obtain a standardized migrant population behavior dataset. Collect entity node information of migrant population, local locations, and social service organizations; mine the relationships between nodes based on standardized migrant population behavior datasets; assign edge weights based on behavior frequency, duration, and stability parameters to obtain a migrant population intelligent agent graph with dynamically weighted associated edges. The system collects the subgraph structure, edge weight time series data, and node association data of the intelligent agent graph. Based on subgraph clustering, weight evolution analysis, and association degree calculation, it obtains three core characteristics for determining permanent residency: spatiotemporal residence, behavioral inertia, and social solidification. Collect historical resident-labeled sample data of the floating population, train the model based on the behavioral characteristics corresponding to regional environmental differences and population attribute differences, and obtain a judgment weight model and a differentiated threshold system that can adaptively adjust feature weights. Based on the three core characteristics of permanent residency—spatiotemporal residence, behavioral inertia, and social solidification—a weighted calculation is performed using the adaptively adjustable characteristic weight model to obtain a comprehensive score for the permanent residency of the floating population. Based on the comprehensive score of the permanent residency of the floating population, the permanent residency determination result and corresponding confidence level of the target floating population are obtained by comparison and identification through the differentiated threshold system.

[0012] In the mobile population permanent residency determination system based on intelligent agent graphs described in this application, the collection of multi-source spatiotemporal behavioral data and social association data of the target mobile population within a preset statistical period is performed on the data according to preset data preprocessing rules to clean, deduplicate, and perform spatiotemporal alignment to obtain a standardized mobile population behavior dataset, specifically: Collect raw spatiotemporal data on the local residence, travel trajectory, and local consumption behavior of the target migrant population within a preset statistical period; Collect raw data on social connections of the target migrant population, including their local employment, public service registration, and local social connections. The original spatiotemporal behavior data and social relational data are processed according to preset data compliance verification and duplicate data removal rules to obtain cleaned and effective data. The cleaned and valid data are time-series aligned based on a unified spatiotemporal coordinate benchmark to obtain a standardized dataset of migrant population behavior with regularized dimensions.

[0013] In the mobile population permanent residency determination system based on agent graph described in this application, the process of collecting entity node information of mobile population, local locations, and social service organizations, mining the relationships between nodes based on a standardized mobile population behavior dataset, and assigning edge weights based on behavior frequency, duration, and stability parameters to obtain a mobile population agent graph with dynamically weighted associated edges, specifically involves: Collect basic attribute data of entity nodes corresponding to individuals of the floating population, local activity venues, and local social service organizations; Based on the standardized migrant population behavior dataset, the association mapping relationship between each entity node is mined. The weights of the associated edges of nodes are quantified based on the frequency of behavior occurrence, cumulative dwell time, and temporal fluctuation stability parameters. By binding each entity node with dynamically assigned associated edge weights, a complete intelligent agent graph of the floating population is obtained.

[0014] Thirdly, this application also provides a computer-readable storage medium, which includes a program for determining the permanent residency of a floating population based on an agent graph. When the program for determining the permanent residency of a floating population based on an agent graph is executed by a processor, it implements the steps of the method for determining the permanent residency of a floating population based on an agent graph as described in any of the preceding claims.

[0015] As can be seen from the above, the method, system, and medium for determining the permanent residency of the floating population based on intelligent agent graphs provided in this application collect multi-source spatiotemporal behavioral data and social correlation data of the floating population and complete standardized preprocessing. An intelligent agent graph with dynamically weighted correlation edges is constructed using the floating population, local locations, and social service institutions as entity nodes. Through graph subgraph clustering, weight temporal evolution, and node correlation calculation, three core features for determining permanent residency—spatiotemporal residence, behavioral inertia, and social solidification—are extracted. An adaptive weighted judgment model and a differentiated threshold system are trained based on historical labeled samples from different regions and population groups. The quantified core features are weighted to obtain a comprehensive permanent residency score. The judgment result and confidence level are output in conjunction with the corresponding threshold. Simultaneously, incremental graph updates and abnormal behavior noise reduction correction optimization logic are added to compensate for the shortcomings of existing technologies, such as thin data dimensions, lack of correlation modeling, fixed judgment thresholds, and severe interference from temporary behaviors. This achieves high-precision automated identification of the permanent residency of the floating population in multi-regional and multi-population scenarios, providing reliable technical support for the refined governance of digital population.

[0016] Other features and advantages of this application will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the embodiments of this application. The objectives and other advantages of this application can be realized through the technical solutions particularly pointed out in the following description and drawings. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart of a method for determining the permanent residency of a floating population based on an agent graph, provided in an embodiment of this application; Figure 2 The flowchart illustrates the process of obtaining three core features for determining permanent residency: spatiotemporal residence, behavioral inertia, and social solidification, based on the agent graph method for determining the permanent residency of the floating population provided in this application embodiment. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0020] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first" and "second" are used only to distinguish descriptions and should not be construed as indicating or implying relative importance. It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions.

[0021] Please refer to Figure 1 , Figure 1 This is a flowchart of a method for determining the permanent residency of a floating population based on an agent graph, as described in some embodiments of this application.

[0022] The first aspect of this invention discloses a method for determining the permanent residency of a floating population based on an intelligent agent graph, which is used in terminal devices such as computers and mobile terminals. This method for determining the permanent residency of a floating population based on an intelligent agent graph includes the following steps: S101. Collect multi-source spatiotemporal behavioral data and social correlation data of the target migrant population within a preset statistical period. Clean, deduplicate, and spatiotemporally align the data according to preset data preprocessing rules to obtain a standardized migrant population behavior dataset. S102. Collect entity node information of migrant population, local locations, and social service organizations. Mine the relationship between nodes based on the standardized migrant population behavior dataset. Assign edge weights according to behavior frequency, duration, and stability parameters to obtain a migrant population intelligent agent graph with dynamically weighted associated edges. S103. Collect the subgraph structure, edge weight time series data and node association data of the intelligent agent graph. Based on subgraph clustering, weight evolution analysis and association degree calculation, obtain three core characteristics of permanent residence: spatiotemporal residence, behavioral inertia and social solidification. S104. Collect historical migrant population permanent resident labeled sample data, train the model according to the behavioral characteristics corresponding to regional environmental differences and population attribute differences, and obtain a judgment weight model and differentiated threshold system that can adaptively adjust feature weights. S105. Based on the three core characteristics of permanent residency—spatiotemporal residence, behavioral inertia, and social solidification—the permanent residency score of the floating population is obtained by performing a weighted calculation through the adaptively adjustable characteristic weight model. S106. Based on the comprehensive score of the permanent residence status of the floating population, the permanent residence status determination result and corresponding confidence level of the target floating population are obtained by comparison and identification through the differentiated threshold system.

[0023] This process involves collecting spatiotemporal behavioral data on the floating population within a 90-day statistical period, including mobile phone signaling usage, offline supermarket consumption, and public transportation use. Simultaneously, it collects social correlation data on local social security enrollment, children's school enrollment, and community residency reporting. The data is then standardized according to rules such as removing missing records, merging duplicate trajectories, and uniformly calibrating coordinates. This ensures compatibility with heterogeneous original information from multiple channels (police, telecom operators, and street offices), eliminating data biases caused by differences in data formats and misaligned collection times, and generating a standardized floating population behavior dataset with unified dimensions. Next, it collects basic information on various entities, including individual floating population members, residential communities, street-front businesses, and government employment service agencies. Based on the standardized dataset, it mines the binding relationships between people and places, and between people and institutions, focusing on long-term residence, stable services, and continuous employment. The weights of the association edges are dynamically calculated using three conventional quantitative indicators: monthly visit frequency, average stay duration per visit, and continuous periodic behavioral fluctuation amplitude. This constructs a real-time updatable intelligent agent graph of the floating population, continuously reflecting changes in the strength of people's local attachment, unlike static association graphs. Finally, it extracts the corresponding local bureaus for each individual within the graph. The system utilizes topological data from subgraphs, continuous weight changes of related edges over the past six months, and deep graph information on the bidirectional correlation density between individuals and local entities. It employs density clustering to delineate high-frequency activity areas, temporal evolution algorithms to capture long-term behavioral patterns, and correlation degree algorithms to quantify the strength of local social binding. This results in the independent output of three non-overlapping core features for determining permanent residency: spatiotemporal residence, behavioral inertia, and social solidification. This effectively filters out invalid behavioral interference from business trips, short-term family visits, and temporary visits. Furthermore, it collects manually labeled population samples from industrial parks, residential communities, and commercial districts over many years. Combined with the behavioral distribution characteristics of different groups such as young migrant workers and families with children, iteratively trains the model to form a determination model and a hierarchical adaptation threshold system that can autonomously adjust feature weights according to local environment and population type. This addresses the shortcomings of traditional unified and fixed determination standards that cannot adapt to diverse urban and rural mobility scenarios. The three types of quantified features obtained from graph analysis are input into the model for weighted fusion calculations. Simultaneously, features deviating from normal values ​​are dulled and corrected to obtain a normalized comprehensive score for permanent residency, avoiding the problem of single-dimensional abnormal data raising or lowering the overall evaluation result.Finally, the comprehensive score is compared and verified against the corresponding regional and population grading thresholds, outputting the confidence level of the resident status determination and the reliability of the representation results. The solution also includes real-time updates to the map based on incremental data, pre-emptive noise reduction and correction optimization logic for abnormal behavior, and monthly integration of newly added migrant population behavior data to iterate the map structure and judgment indicators, further reducing the interference of short-term travel and temporary stay data on the final determination. The entire chain relies on the intelligent agent map to integrate spatiotemporal trajectory and social connection information, coupled with adaptive dynamic judgment logic. Compared with traditional identification methods that rely solely on single mobile phone signaling duration and static population registration information, it significantly improves the adaptability and objectivity of migrant population resident status identification, and can support the implementation of digital grassroots population governance services such as residence permit qualification verification, total population assessment of the jurisdiction, and balanced allocation of public resources such as education and medical care.

[0024] According to an embodiment of the present invention, the multi-source spatiotemporal behavioral data and social correlation data of the target migrant population within a preset statistical period are collected, and the data are cleaned, deduplicated, and spatiotemporally aligned according to preset data preprocessing rules to obtain a standardized migrant population behavior dataset, specifically: Collect raw spatiotemporal data on the local residence, travel trajectory, and local consumption behavior of the target migrant population within a preset statistical period; Collect raw data on social connections of the target migrant population, including their local employment, public service registration, and local social connections. The original spatiotemporal behavior data and social relational data are processed according to preset data compliance verification and duplicate data removal rules to obtain cleaned and effective data. The cleaned and valid data are time-series aligned based on a unified spatiotemporal coordinate benchmark to obtain a standardized dataset of migrant population behavior with regularized dimensions.

[0025] The process involves collecting raw data on the spatiotemporal behavior of migrant populations within a 90-day standard statistical period, including community residency signals, commuting trajectories, and offline consumption records at supermarkets and stores. Simultaneously, it collects raw data on social connections related to local employment, public services, and social interactions, such as enterprise social security payment records, community preschool education registrations, and local neighborhood social reporting. Invalid records exceeding the jurisdiction or collection period are removed using compliance verification rules, and duplicate trajectory entries generated simultaneously by multiple devices for the same person are merged using duplicate data removal rules. This results in cleaned, redundant data. Finally, all cleaned data is aligned and calibrated using a nationally unified geographic coding and time stamp to eliminate deviations caused by data collection time offsets and inconsistent coordinate standards between operators and government platforms. The final result is a standardized migrant population behavior dataset with unified fields, synchronized time, and direct applicability for correlation mining. This dataset is compatible with heterogeneous data sources from multiple departments, reducing errors in subsequent map construction and feature extraction processes caused by defects in the raw data.

[0026] According to an embodiment of the present invention, the process of collecting entity node information of migrant population, local locations, and social service organizations, mining relationships between nodes based on a standardized migrant population behavior dataset, and assigning edge weights based on behavior frequency, duration, and stability parameters to obtain a migrant population agent graph with dynamically weighted associated edges, specifically involves: Collect basic attribute data of entity nodes corresponding to individuals of the floating population, local activity venues, and local social service organizations; Based on the standardized migrant population behavior dataset, the association mapping relationship between each entity node is mined. The weights of the associated edges of nodes are quantified based on the frequency of behavior occurrence, cumulative dwell time, and temporal fluctuation stability parameters. By binding each entity node with dynamically assigned associated edge weights, a complete intelligent agent graph of the floating population is obtained.

[0027] This process involves collecting basic attribute data on the physical nodes of migrant population identity identifiers, residential communities, street-front shops, street government service centers, and industrial parks, as well as their corresponding activity locations and social service organizations. Based on a standardized migrant population behavior dataset previously compiled, the process uncovers long-term activity and stable business relationships between personnel nodes and residential location nodes, and between personnel nodes and employment organization nodes. Combining three quantitative parameters—monthly behavior frequency, cumulative stay duration within a cycle, and the stability of behavior temporal fluctuations over consecutive periods—the process numerically calculates the relationships between nodes. The dynamically weighted data obtained in real-time is then bound to the relationships between the corresponding physical nodes, ultimately generating a complete and quantifiable migrant population intelligent agent graph. This graph can intuitively quantify the strength of dependence between individuals and local entities, distinguishing it from static relationship graphs without weight labels. It provides a quantifiable graph structure foundation for subsequent extraction of permanent residency determination features.

[0028] Please refer to Figure 2 , Figure 2 This is a flowchart illustrating the process of obtaining the video playback frame position corresponding to a no-whistle dispute scene in some embodiments of this application, based on the agent graph-based method for determining the permanent residency of a floating population. According to embodiments of the present invention, the process involves collecting the subgraph structure, edge weight temporal data, and node association data of the agent graph, and obtaining three core features for determining permanent residency—spatiotemporal residence, behavioral inertia, and social solidification—based on subgraph clustering, weight evolution analysis, and association degree calculation. Specifically: S201. Collect the subgraph topology and node association density data of the mobile population intelligent agent graph; S202. Collect the temporal weight evolution data of the associated edges of the mobile population intelligent agent graph and record the weight fluctuation change status within the period. S203. Based on subgraph clustering, temporal evolution analysis, and node correlation calculation methods, quantify the multidimensional features of the graph respectively; S204. Based on the quantification results of the features, the three core judgment features are classified and integrated to obtain spatiotemporal dwell, behavioral inertia, and social solidification.

[0029] This process involves collecting data on the topological structure of local subgraphs corresponding to individual migrants, the node association density of their locations and service institutions, such as the distribution structure of local subgraphs corresponding to daily residence, consumption, and commuting, and the degree of association between migrants and local entities. Secondly, it collects time-series weight evolution data of all associated edges in the graph over a statistical period of several months, continuously recording the weight increases, decreases, fluctuations, and steady-state changes of residency associations, employment associations, and public service associations. Subsequently, it aggregates and divides the activity areas of the migrant population using a subgraph clustering algorithm, captures long-term behavioral stability patterns through time-series evolution analysis, and quantifies the degree of social binding through node association calculation, achieving multi-dimensional quantification of graph structure, time-series changes, and association strength. Finally, it categorizes and integrates the various quantification results to form spatiotemporal residency characteristics representing the stability of residency, behavioral inertia characteristics representing the continuity of long-term behavior, and social solidification characteristics representing the degree of residency attachment. This transforms the original graph structure into core features for determining permanent residency, effectively removing invalid feature interference from temporary visits and short-term stays, and providing differentiated and highly distinguishable core feature support for subsequent accurate quantification of the permanent residency attributes of the migrant population.

[0030] According to an embodiment of the present invention, the step of collecting historical migrant population permanent resident labeled sample data, training a model based on behavioral characteristic patterns corresponding to regional environmental differences and population attribute differences, and obtaining a judgment weight model and a differentiated threshold system that can adaptively adjust feature weights are specifically as follows: Collect manually labeled historical permanent residence data of the floating population in different regions and functional scenarios; Collect behavioral characteristic distribution data of migrant population samples with different age structures and employment attributes; The initial model parameters are iteratively trained based on the differences in behavioral characteristics corresponding to regional environment and population attributes. Based on the converged model parameters, an adaptive weight adjustment model and a differentiated threshold system for matching scenarios are obtained.

[0031] This process involves collecting historical manually labeled sample data of the floating population in various local scenarios, specifically covering permanent and temporary resident samples from different functional areas such as urban industrial parks, old residential areas, commercial districts, and suburban migrant worker areas, thus covering diverse local environmental scenarios. Simultaneously, it collects behavioral characteristic distribution data of the floating population with different attributes, including spectral feature data of different age groups and employment types, such as working-age groups, family-based groups, flexible employment groups, and short-term employment groups, enriching the dimensions of the model training samples and improving the sample coverage. This step relies on massive amounts of differentiated sample data to mine the differences in population activity, resource allocation, and behavioral characteristics corresponding to the different migration habits and local binding patterns of different groups in different regions. It continuously iterates and optimizes the feature matching parameters and judgment parameters of the initial model, eliminating the limitations of training with samples from a single scenario or a single population. An adaptive weight adjustment model is constructed based on the optimal model parameters after training convergence. It can dynamically match the weight ratio of the three core judgment features according to the scene and population characteristics, and generate a hierarchical and differentiated judgment threshold system that is adapted to different local scenes and different population types. This avoids the use of traditional unified and fixed judgment standards, effectively improves the scene generalization ability and population adaptability of the mobile population permanent residence judgment model, makes the judgment standard conform to the real population flow pattern, and greatly reduces the judgment error in differentiated scenarios.

[0032] According to an embodiment of the present invention, the comprehensive score of the permanent residency of the floating population is obtained by weighting the three core characteristics of permanent residency—spatiotemporal residence, behavioral inertia, and social solidification—through a weighted calculation using a judgment weight model with adaptively adjustable feature weights. Specifically: Collect quantitative values ​​of the three core judgment features: spatiotemporal dwell time, behavioral inertia, and social solidification; Collect the real-time weight configuration parameters of the trained adaptive decision weight model; The three core features are multidimensionally weighted and fused according to the real-time weight configuration parameters of the adaptive judgment weight model to obtain the initial score data; The initial scoring data is processed according to a preset normalization correction rule to obtain a standardized comprehensive score for the permanent residence of the floating population.

[0033] The process involves collecting specific quantitative values ​​of three core judgment features—spatiotemporal residence, behavioral inertia, and social solidification—obtained through graph mining. These values ​​cover multidimensional features related to the stability of a person's place of residence, the regularity of long-term behavior, and the strength of social bonds. Simultaneously, real-time weight configuration parameters of the trained adaptive judgment weight model are collected under the current scenario and population adaptation. Compared to fixed weight ratios, these weight parameters can be dynamically adjusted based on the functional attributes of the region and the attributes of the floating population. Subsequently, based on the real-time weight configuration parameters output by the model, a multidimensional weighted fusion operation is performed on the three independent core judgment features. This integrates multi-level feature information of spatiotemporal behavior and social association, calculating an irregular initial score for permanent residency, thus achieving effective fusion of multi-source features. Finally, based on the preset normalization correction rules, the initial scoring data is uniformly corrected, range scaled, and outlier passivation is performed to eliminate scoring bias caused by differences in the dimensions of different features and local extreme outliers. The output is a standardized comprehensive score of the permanent residence status of the floating population with uniform range, regular dimensions, and horizontal comparability, which effectively ensures the stability, objectivity, and comparability of the scoring results and provides reliable quantitative data support for the subsequent accurate determination of the permanent residence status of the floating population.

[0034] According to an embodiment of the present invention, it further includes: Collect incremental behavioral data and social relationship update data of the target floating population within the judgment period; The node and associated edge weights of the mobile population intelligent agent graph are updated in real time based on the time-series incremental data. The core judgment features and the comprehensive score of permanent residency are simultaneously corrected based on the dynamic update results of the map. The threshold comparison is completed based on the comprehensive score after real-time correction, and the dynamic updated permanent resident determination result is obtained.

[0035] This involves collecting incremental behavioral data and social association updates monthly within the target migrant population determination period. This includes real-time updates of newly added local residency trajectories, monthly travel and consumption behaviors, new social security enrollment records, and public service registration information, continuously updating the latest local behavior and social binding status of individuals. This step uses time-series incremental data to iteratively update the migrant population intelligent agent graph in real time. For newly added social binding relationships, corresponding entity association nodes are added. Simultaneously, the weights of existing association edges are dynamically adjusted based on the latest behavior frequency and residency duration data, avoiding the static and fixed construction mode of traditional graphs and ensuring that the graph structure and association strength remain consistent with the latest local status of individuals. Based on this, according to the dynamically updated graph topology and time-series weight parameters, the quantitative values ​​of three core determination features—spatiotemporal residency, behavioral inertia, and social solidification—are recalculated and corrected synchronously, and the comprehensive score for migrant population permanent residency is updated in conjunction with this, avoiding the determination lag caused by long-term fixed features and scores. Finally, based on the real-time corrected permanent residence comprehensive score, the system matches the corresponding scenario-differentiated thresholds to complete the comparison and identification, and outputs dynamically updated permanent residence judgment results and confidence levels. This effectively adapts to scenarios of status changes such as seasonal employment, long-term settlement, and new local social relationships of the floating population, continuously improving the timeliness and accuracy of the identification of the permanent residence status of the population at all times, and ensuring that the judgment results are consistent with the actual local residence patterns of the personnel.

[0036] According to an embodiment of the present invention, it further includes: Collect disruptive behavioral data on short-term stays and temporary travel of the floating population in the intelligent agent graph; Invalid interference data samples are filtered and isolated according to preset abnormal behavior identification rules; The graph topology and temporal weight evolution trend are corrected based on the anti-interference optimization algorithm; Based on the optimized map data, three types of core judgment features are calculated and extracted to complete feature denoising correction.

[0037] The process involves continuously collecting non-steady-state disruptive behavioral data on the mobile population within the intelligent entity graph, including short-term stays, business trips, temporary visits, and brief transits. This type of behavioral data has a short dwell period and high randomness, easily confusing with resident behavior characteristics and leading to misjudgments. Subsequently, based on pre-set abnormal behavior identification rules, the entire graph's behavioral data is subjected to stratified screening to accurately identify invalid disruptive data samples with single short-term stays, low-frequency visits, and no social connections. These disruptive samples are then isolated, marked, and stripped, eliminating invalid noise information at the data source. Furthermore, a dedicated anti-interference optimization algorithm is used to globally correct the graph's topology and temporal weight evolution trends, weakening the impact of temporary behaviors on the subgraph's clustering structure and suppressing the interference of short-term fluctuations in data on the evolution trajectory of associated edge weights, ensuring that the overall graph structure conforms to the long-term resident behavior patterns of individuals. Finally, based on the high-quality map data after noise reduction and correction, the three core judgment features of spatiotemporal residence, behavioral inertia, and social solidification were recalculated and extracted. This completed the precise noise reduction and correction of the feature dimensions, effectively eliminating the feature distortion problem caused by occasional behavior, and greatly improving the stability and distinguishability of the core judgment features. This made the final permanent residence judgment result more consistent with the actual long-term local residence status of the floating population.

[0038] The second aspect of the present invention also discloses a system for determining the permanent residency of a floating population based on an agent graph, comprising a memory and a processor. The memory includes a program for determining the permanent residency of a floating population based on an agent graph. When the program for determining the permanent residency of a floating population based on an agent graph is executed by the processor, it implements the steps of the method for determining the permanent residency of a floating population based on an agent graph as described in any of the preceding claims.

[0039] A third aspect of the present invention provides a computer-readable storage medium comprising a method program for determining the permanent residency of a floating population based on an agent graph. When the method program for determining the permanent residency of a floating population based on an agent graph is executed by a processor, it implements the steps of the method for determining the permanent residency of a floating population based on an agent graph as described in any of the preceding claims.

[0040] This invention discloses a method, system, and medium for determining the permanent residency of migrant populations based on intelligent agent graphs. First, it collects multi-source spatiotemporal behavioral raw data on migrant populations within a preset statistical period, including their place of residence, travel trajectories, and local consumption, as well as social association raw data such as local employment binding, public service registration, and local social binding. Through standardized preprocessing methods such as compliance verification, redundancy removal, and spatiotemporal alignment, a unified and well-organized migrant population behavior dataset is obtained, solving the analytical bias problems caused by the disorder and temporal misalignment of multi-source heterogeneous data. Based on this, it aggregates various types of data, including individual migrant populations, local activity venues, and social service institutions. By analyzing the attribute information of entity nodes and leveraging standardized behavioral datasets to mine the association mapping relationships between different entities, and dynamically quantifying the weights of association edges based on parameters such as behavior frequency, cumulative residence time, and temporal fluctuation stability, a dynamic weighted intelligent agent graph of the floating population is constructed to achieve structured and quantitative analysis of personnel's local behavior and social dependence. Subsequently, by collecting data on the topology of the graph subgraphs, node association density, and temporal weight evolution, and combining subgraph clustering, weight temporal evolution analysis, and association quantification calculation methods, layered extraction is performed to extract data that can respectively characterize the stability of local residence, long-term behavioral regularity, and local social dependence. This invention establishes three independent and complementary core features for determining permanent residency: spatiotemporal residence, behavioral inertia, and social solidification. It also identifies and isolates interfering behavioral data such as short-term stays and temporary travel, and utilizes anti-interference optimization algorithms to correct the topology and weight evolution trends of the geographic map, achieving noise reduction and correction of core features and effectively avoiding judgment distortion caused by temporary and sporadic behaviors. Furthermore, this invention collects historical permanent residency annotation sample data from different geographical locations and population attributes, and iteratively trains model parameters by combining regional environmental and behavioral differences to construct a judgment model that can adaptively adjust feature weights and adapt to population differences across multiple scenarios. The invention employs a differentiated threshold system, overcoming the shortcomings of traditional fixed threshold judgment methods in terms of poor scenario adaptability. In the judgment stage, the invention collects three types of core feature quantification values ​​and real-time model weight parameters to complete multi-dimensional feature weighted fusion calculation and outputs a standardized resident comprehensive score by combining normalization correction rules. It relies on the differentiated threshold system to complete score comparison and identification, outputting resident judgment results and corresponding confidence levels. At the same time, it supports periodic collection of incremental behavior and social relationship update data, real-time iterative update of graph nodes and weight parameters, synchronous correction of judgment features and comprehensive scores, and realization of dynamic time-series update of judgment results.This invention also features a dedicated judgment system and computer-readable storage medium. Through a hardware and software collaborative architecture, it stably executes the entire judgment logic. This comprehensive technical solution completely transforms the traditional, crude method of identifying permanent residents based solely on single residency duration and static registration data. Leveraging multiple innovative mechanisms such as multi-dimensional modeling of intelligent agent graphs, adaptive differentiated judgment, anti-interference noise reduction, and time-series dynamic updates, it significantly improves the accuracy, stability, and scenario generalization capability of determining the permanent residency of the floating population. It can adapt to diverse population governance scenarios such as urban and rural residential areas, industrial parks, and commercial districts, providing reliable and efficient intelligent technical support for refined government governance work such as grassroots digital population statistics, residence permit qualification verification, and precise allocation of public service resources.

[0041] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0042] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0043] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0044] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory, random access memory, magnetic disks, or optical disks.

[0045] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

Claims

1. A method for determining the permanent residency of a floating population based on agent graphs, characterized in that, Includes the following steps: Collect multi-source spatiotemporal behavioral data and social correlation data of the target migrant population within a preset statistical period, and clean, deduplicate, and spatiotemporally align the data according to preset data preprocessing rules to obtain a standardized migrant population behavior dataset. Collect entity node information of migrant population, local locations, and social service organizations; mine the relationships between nodes based on standardized migrant population behavior datasets; assign edge weights based on behavior frequency, duration, and stability parameters to obtain a migrant population intelligent agent graph with dynamically weighted associated edges. The system collects the subgraph structure, edge weight time series data, and node association data of the intelligent agent graph. Based on subgraph clustering, weight evolution analysis, and association degree calculation, it obtains three core characteristics for determining permanent residency: spatiotemporal residence, behavioral inertia, and social solidification. Collect historical resident-labeled sample data of the floating population, train the model based on the behavioral characteristics corresponding to regional environmental differences and population attribute differences, and obtain a judgment weight model and a differentiated threshold system that can adaptively adjust feature weights. Based on the three core characteristics of permanent residency—spatiotemporal residence, behavioral inertia, and social solidification—a weighted calculation is performed using the adaptively adjustable characteristic weight model to obtain a comprehensive score for the permanent residency of the floating population. Based on the comprehensive score of the permanent residency of the floating population, the permanent residency determination result and corresponding confidence level of the target floating population are obtained by comparison and identification through the differentiated threshold system.

2. The method for determining the permanent residency of the floating population based on agent graphs according to claim 1, characterized in that, The multi-source spatiotemporal behavioral data and social correlation data of the target migrant population within a preset statistical period are collected. The data is then cleaned, deduplicated, and spatiotemporally aligned according to preset data preprocessing rules to obtain a standardized migrant population behavior dataset, specifically: Collect raw spatiotemporal data on the local residence, travel trajectory, and local consumption behavior of the target migrant population within a preset statistical period; Collect raw data on social connections of the target migrant population, including their local employment, public service registration, and local social connections. The original spatiotemporal behavior data and social relational data are processed according to preset data compliance verification and duplicate data removal rules to obtain cleaned and effective data. The cleaned and valid data are time-series aligned based on a unified spatiotemporal coordinate benchmark to obtain a standardized dataset of migrant population behavior with regularized dimensions.

3. The method for determining the permanent residency of the floating population based on agent graphs according to claim 1, characterized in that, The process involves collecting entity node information from the migrant population, local locations, and social service organizations; mining relationships between nodes based on a standardized migrant population behavior dataset; and assigning edge weights based on behavior frequency, duration, and stability parameters to obtain a migrant population agent graph with dynamically weighted edges. Specifically: Collect basic attribute data of entity nodes corresponding to individuals of the floating population, local activity venues, and local social service organizations; Based on the standardized migrant population behavior dataset, the association mapping relationship between each entity node is mined. The weights of the associated edges of nodes are quantified based on the frequency of behavior occurrence, cumulative dwell time, and temporal fluctuation stability parameters. By binding each entity node with dynamically assigned associated edge weights, a complete intelligent agent graph of the floating population is obtained.

4. The method for determining the permanent residency of the floating population based on agent graphs according to claim 1, characterized in that, The collected intelligent agent graph's subgraph structure, edge weight time-series data, and node association data are used to obtain three core characteristics for determining permanent residency: spatiotemporal residence, behavioral inertia, and social solidification, based on subgraph clustering, weight evolution analysis, and association degree calculation methods. Specifically: Collect the subgraph topology and node association density data of the intelligent agent graph of the floating population; Collect time-series weight evolution data of the associated edges of the mobile population intelligent agent graph, and record the weight fluctuation and change status within the period; Based on subgraph clustering, temporal evolution analysis, and node correlation calculation methods, the multidimensional features of the graph are quantified respectively. Based on the quantification results of the features, three core judgment features were obtained: spatiotemporal dwell time, behavioral inertia, and social solidification.

5. The method for determining the permanent residency of the floating population based on agent graphs according to claim 1, characterized in that, The collected historical migrant population permanent resident labeled sample data is used to train a model based on behavioral characteristic patterns corresponding to regional environmental differences and population attribute differences. This results in an adaptively adjustable judgment weight model and a differentiated threshold system, specifically: Collect manually labeled historical permanent residence data of the floating population in different regions and functional scenarios; Collect behavioral characteristic distribution data of migrant population samples with different age structures and employment attributes; The initial model parameters are iteratively trained based on the differences in behavioral characteristics corresponding to regional environment and population attributes. Based on the converged model parameters, an adaptive weight adjustment model and a differentiated threshold system for matching scenarios are obtained.

6. The method for determining the permanent residency of a floating population based on an agent graph according to claim 1, characterized in that, The process involves weighting the three core characteristics of permanent residency—spatiotemporal residence, behavioral inertia, and social solidification—using a weighted calculation model with adaptively adjustable feature weights to obtain a comprehensive score for the permanent residency of the floating population. Specifically: Collect quantitative values ​​of the three core judgment features: spatiotemporal dwell time, behavioral inertia, and social solidification; Collect the real-time weight configuration parameters of the trained adaptive decision weight model; The three core features are multidimensionally weighted and fused according to the real-time weight configuration parameters of the adaptive judgment weight model to obtain the initial score data; The initial scoring data is processed according to a preset normalization correction rule to obtain a standardized comprehensive score for the permanent residence of the floating population.

7. A system for determining the permanent residency of a floating population based on an intelligent agent graph, characterized in that, The system includes a memory and a processor. The memory contains a program for determining the permanent residency of the floating population based on an agent graph. When the processor executes the program for determining the permanent residency of the floating population based on an agent graph, it performs the following steps: Collect multi-source spatiotemporal behavioral data and social correlation data of the target migrant population within a preset statistical period, and clean, deduplicate, and spatiotemporally align the data according to preset data preprocessing rules to obtain a standardized migrant population behavior dataset. Collect entity node information of migrant population, local locations, and social service organizations; mine the relationships between nodes based on standardized migrant population behavior datasets; assign edge weights based on behavior frequency, duration, and stability parameters to obtain a migrant population intelligent agent graph with dynamically weighted associated edges. The system collects the subgraph structure, edge weight time series data, and node association data of the intelligent agent graph. Based on subgraph clustering, weight evolution analysis, and association degree calculation, it obtains three core characteristics for determining permanent residency: spatiotemporal residence, behavioral inertia, and social solidification. Collect historical resident-labeled sample data of the floating population, train the model based on the behavioral characteristics corresponding to regional environmental differences and population attribute differences, and obtain a judgment weight model and a differentiated threshold system that can adaptively adjust feature weights. Based on the three core characteristics of permanent residency—spatiotemporal residence, behavioral inertia, and social solidification—a weighted calculation is performed using the adaptively adjustable characteristic weight model to obtain a comprehensive score for the permanent residency of the floating population. Based on the comprehensive score of the permanent residency of the floating population, the permanent residency determination result and corresponding confidence level of the target floating population are obtained by comparison and identification through the differentiated threshold system.

8. The system for determining the permanent residency of a floating population based on an intelligent agent graph according to claim 7, characterized in that, The multi-source spatiotemporal behavioral data and social correlation data of the target migrant population within a preset statistical period are collected. The data is then cleaned, deduplicated, and spatiotemporally aligned according to preset data preprocessing rules to obtain a standardized migrant population behavior dataset, specifically: Collect raw spatiotemporal data on the local residence, travel trajectory, and local consumption behavior of the target migrant population within a preset statistical period; Collect raw data on social connections of the target migrant population, including their local employment, public service registration, and local social connections. The original spatiotemporal behavior data and social relational data are processed according to preset data compliance verification and duplicate data removal rules to obtain cleaned and effective data. The cleaned and valid data are time-series aligned based on a unified spatiotemporal coordinate benchmark to obtain a standardized dataset of migrant population behavior with regularized dimensions.

9. The system for determining the permanent residency of a floating population based on an intelligent agent graph according to claim 7, characterized in that, The process involves collecting entity node information from the migrant population, local locations, and social service organizations; mining relationships between nodes based on a standardized migrant population behavior dataset; and assigning edge weights based on behavior frequency, duration, and stability parameters to obtain a migrant population agent graph with dynamically weighted edges. Specifically: Collect basic attribute data of entity nodes corresponding to individuals of the floating population, local activity venues, and local social service organizations; Based on the standardized migrant population behavior dataset, the association mapping relationship between each entity node is mined. The weights of the associated edges of nodes are quantified based on the frequency of behavior occurrence, cumulative dwell time, and temporal fluctuation stability parameters. By binding each entity node with dynamically assigned associated edge weights, a complete intelligent agent graph of the floating population is obtained.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a method program for determining the permanent residency of a floating population based on an agent graph. When the method program for determining the permanent residency of a floating population based on an agent graph is executed by a processor, it implements the steps of the method for determining the permanent residency of a floating population based on an agent graph as described in any one of claims 1 to 6.

Images