Property payment management system based on big data

Abstract

The invention discloses a property payment management system based on big data, and relates to the technical field of property payment management, and the system carries out the real-time monitoring and analysis of the payment behavior of a resident through a random forest and a graph neural network, is not limited to a static rule, can dynamically recognize the complex abnormal behavior, and achieves the real-time monitoring and analysis of the payment behavior of the resident. Risk levels of residents are automatically output through the model, high-risk residents are marked in real time, dependence on manual auditing is reduced, meanwhile, accuracy and timeliness of recognition are improved, long-term or intermittent drawback problems of the residents are effectively solved, payment confirmation and overdue fine calculation are automatically executed through an intelligent contract, automatic updating of payment management is achieved, and user experience is improved. According to the payment capability, the defaulted amount and the performance condition of the resident, a dynamically adjusted personalized payment scheme is generated, intelligent reminding is automatically pushed before the payment day, the manual confirmation step is reduced, the state updating speed is increased, and the payment period and the reminding frequency are adjusted according to the newest behavior of the resident.

Description

Technical Field

[0001] The present invention relates to the technical field of property payment management, and in particular to a property payment management system based on big data. Background Art

[0002] With the continuous development of technology in the era of big data, it has played a positive role in promoting data processing processes and related structures, greatly changing people's lives. The community property management system supported by the big data background can better serve property management, comprehensively improve the quality of property services, and bring greater benefits to property companies.

[0003] In the existing property payment management system based on big data, the monitoring and analysis methods of residents' payment behavior are mostly based on preset data rules and historical data statistics for anomaly identification, identifying users' long-term arrears or large unpaid problems, and using preset rules to detect abnormal payment behavior. Although it is easy to implement, the rule settings are relatively simple and difficult to change, and cannot cope with complex abnormal situations, such as users' long-term arrears or indirect arrears of property fees. For such abnormal behaviors, further manual confirmation and processing are required, which is manpower-consuming and there are problems of information lag and untimely communication. Therefore, there is an urgent need for a property payment management system based on big data to solve such problems. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a property payment management system based on big data to solve the problems that the rule engine adopted by the traditional solution can only process predefined rules, cannot adapt to the dynamic changes of user behavior, lacks the ability to respond to newly emerging abnormal behaviors, is prone to errors and omissions, and has a slow response speed.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] On the one hand, the present invention provides a property payment management system based on big data, which includes:

[0008] The property payment data collection module is responsible for collecting and managing property payment related data, including property contract terms, payment rules, resident payment records and contract performance information;

[0009] The resident labeling and classification module analyzes the payment data of residents, identifies abnormal payment behaviors, outputs abnormal detection results, and classifies the risk level of residents into high, medium, and low;

[0010] The risk assessment module conducts risk assessment and classification of residents based on anomaly detection results and dynamically updates the risk level of residents;

[0011] The high-risk resident payment management module provides personalized payment reminders and management strategies for high-risk residents;

[0012] Smart contract execution and payment automation module converts property contract terms, payment rules and anomaly detection results into smart contracts, and is responsible for all automatically executed operations in the property payment management system, including payment confirmation, overdue processing and fee adjustment;

[0013] The early warning intervention module monitors the payment behavior and feedback of high-risk residents and issues early warnings to residents who may incur financial risks.

[0014] On the other hand, the present invention provides a property payment management method based on big data, comprising:

[0015] Step S1, payment data collection and abnormal behavior identification, collects the payment data of the residents, integrates and preprocesses it, uses the random forest model to preliminarily analyze the payment records of the residents, and outputs the abnormal detection results, that is, the abnormal behavior of the residents.

[0016] Based on abnormal behavior identification, a resident behavior graph is constructed, and the resident's payment mode is combined with the contract performance to output the resident's risk level, which is divided into high, medium and low, and high-risk residents are marked, that is, residents who frequently default or owe large amounts of fees;

[0017] Step S2, high-risk resident management, based on the risk level output in step S1, tracking the payment behavior of high-risk residents, and pushing smart reminders and personalized payment plans to them;

[0018] Smart reminders and personalized payment plans include:

[0019] Before the payment date, payment reminders will be automatically pushed to remind residents to pay on time.

[0020] According to the amount of arrears and contract performance of the residents, the payment ability and behavior characteristics of the residents are analyzed, and a personalized plan, namely the basic payment plan and installment suggestions, is generated and pushed to the residents. The plan is dynamically adjusted based on the amount of arrears and contract performance, reflecting the latest risk status of the residents in real time;

[0021] Step S3, smart contract execution and automated payment management, converting payment rules, contract terms and anomaly detection results into smart contracts, automatically executing payment confirmation, overdue penalty calculation and fee adjustment,

[0022] The smart contract adjusts the payment period and reminder frequency for high-risk residents based on the personalized plan in step S2, and automatically matches the residents' payment ability and behavioral characteristics analyzed in step S2.

[0023] Furthermore, in step S1, abnormal behaviors include frequent overdue payments, large arrears, and intermittent arrears.

[0024] In step S1, the payment data includes property contract terms, payment rules, resident payment records and contract performance information.

[0025] Furthermore, in step S1, the household payment records are analyzed and the abnormality detection results are outputted in the following manner:

[0026] Preprocess the collected payment data. Suppose the payment data set is X = {x1, x2, ..., x n}, where each x i Represents a household’s feature vector, including historical payment records, overdue times, and contract performance. The payment behavior features are constructed as the input vector x i =[f1,f2,…,f m ],in,

[0027] Among them, f1 represents the relative overdue frequency of the household, Odds represents the total number of overdue days of the household in a cycle, and Tpp represents the total number of payment days stipulated in the contract.

[0028] Among them, f2 represents the relative arrears amount of the household, Ao represents the total amount of unpaid amount of the household in the period, and Ap represents the amount payable calculated according to the contract terms.

[0029] f3 represents the contract performance score, which is based on the performance of the contract terms and ranges from 0 to 1.

[0030] Building a Random Forest Model Contains T decision trees, and the output of each tree is The model output is: in, represents the predicted anomaly score, which measures the abnormal payment behavior of the resident. Indicates The output of a decision tree is

[0031] Define the threshold τ to identify abnormal behavior, if If , it is abnormal behavior, otherwise it is normal behavior. τ is set based on the statistical analysis of historical data.

[0032] Furthermore, in step S1, a household behavior graph is constructed based on the anomaly detection results, and the household risk level is output in combination with the payment mode and contract performance:

[0033] Extract key behavioral features from anomaly detection results, including:

[0034] f freq_delay, represents the frequency of overdue payments of residents in multiple payment cycles,

[0035] f large_debt , indicating the accumulated large amount of arrears of the household,

[0036] f irreg_arrears , indicating the frequency and pattern of episodic arrears of households,

[0037] f contract_compliance represents the household's contract performance score,

[0038] Construct a graph structure, G = (V, E), where V represents the node set, E represents the edge set, and each node v i ∈V represents a household, and its feature vector h is defined i for:

[0039] h i =[f freq_delay ,f large_debt ,f irreg_arrears ,f contract_compliance ], where h i The feature vector representing household i includes overdue frequency, large amount of arrears, intermittent arrears and compliance score;

[0040] Connecting resident nodes with similar payment behaviors, the edge weight is w ij represents the similarity of the behaviors between two residents and is defined as:

[0041] Among them, w ij represents the behavioral similarity weight between residents i and j, |h i -h j | represents the Euclidean distance between eigenvectors, σ represents the smoothing parameter;

[0042] Perform graph convolution on the constructed household behavior graph to obtain the comprehensive behavior characteristics of each household:

[0043] Among them, h′ i represents the feature vector of household i after graph convolution update, represents the set of neighbor nodes of household i, W represents the weight matrix of the graph convolutional layer, and σ represents the activation function;

[0044] The updated feature vector h′ i Input the risk assessment function to obtain the resident’s risk score ri:

[0045] r i =α1f′ freq_delay +α2f′ large_debt +α3f′ irreg_arrears +α4f′contract_compliance , where r i represents the risk score of household i, α1, α2, α3, α4 represent the characteristic weight coefficients;

[0046] According to the risk score r i , classifying households into high, medium, and low risk levels:

[0047] If r i >θ high , it is classified as high risk level.

[0048] If θ low <r i ≤θ high , it is classified as medium risk level.

[0049] If r i ≤θ low , it is classified as low risk level.

[0050] Among them, θ high represents the high risk threshold, θ low Indicates a low risk threshold.

[0051] Furthermore, in step S2, based on the risk level output in step S1, the payment behavior of high-risk residents is tracked, and intelligent reminders and personalized payment plans are pushed to them:

[0052] Identify the high-risk household set from the household risk level output in step S1, set where r i represents the risk score of household i, θ high is a high risk threshold,

[0053] Each high-risk resident Track their payment behavior in real time and build a dynamic feature update model:

[0054] in, represents the behavioral feature vector of household i at time t+1, represents the behavioral feature vector of household i at time t, represents the latest observed behavioral feature vector of household i, including the latest payment behavior and contract performance, and α represents the time decay factor;

[0055] According to the updated feature vector, analyze the behavior trend of the residents and predict their payment ability, and output the payment ability score c i :

[0056] Among them, c irepresents the payment ability score of household i, ω1, ω2, ω3, ω4 represent the characteristic weights, Indicates the updated overdue frequency. Indicates the updated large amount of arrears feature. represents the updated intermittent delinquency characteristics, represents the updated contract performance score;

[0057] Based on the ability to pay rating c i , generate a personalized payment plan and set the basic payment amount p i and staging recommendations i :

[0058] p i =γ·Ao, where p i represents the basic payment amount, which is the initial payment amount set by high-risk households based on their actual payment ability; γ represents the payment ratio coefficient, which depends on the household’s risk level and payment ability;

[0059] Among them, s i represents the installment payment amount, which is the remaining arrears divided equally into the next η periods, where η represents the number of installments, and is dynamically adjusted based on the household's risk level and payment ability;

[0060] Set up smart reminder strategies based on residents’ behavioral trends and personalized payment plans:

[0061] Among them, Rt represents the specific time of pushing payment reminders in advance, T represents the payment date, and δ represents the reminder advance parameter, which depends on the payment ability score c i , the lower the score, the earlier the reminder.

[0062] Furthermore, in step S2, based on the amount of the household's arrears and contract performance, the household's payment ability and behavior characteristics are analyzed to generate a personalized plan:

[0063] For each household i, construct the affordability scoring model c i :

[0064] Among them, c i represents the payment ability score of household i, φ1, φ2, φ3 represent the characteristic weight coefficients, inc i represents the monthly income of household i, Ao i represents the total amount of arrears currently owed by household i, Noop i represents the number of overdue payments of household i in the most recent cycle, r i represents the contract performance score of household i,

[0065] Assess the payment behavior characteristics of households and calculate the risk adjustment coefficient ψ i :

[0066] Among them, ψ i represents the risk adjustment coefficient of household i, κ1, κ2, κ3 represent the characteristic weights, Ap i Tdo is the total amount of fees that household i should pay according to the contract. i represents the cumulative number of overdue days for household i, Cp i represents the length of the payment contract period of household i, f Fd Characteristics indicating that households frequently default over multiple periods;

[0067] Based on the ability to pay scorec i and the risk adjustment factor ψ i , generating the household's basic payment amount p i and staging recommendations i :

[0068] Among them, p i represents the basic payment amount of household i, that is, the first payment amount calculated in a personalized manner, μ represents the basic payment ratio coefficient, and MiniPs represents the minimum payment amount standard;

[0069] Among them, s i represents the installment payment amount of household i, the remaining arrears amount is evenly distributed, ξ represents the installment adjustment factor, and the risk coefficient ψ i Dynamically adjust the number of installments;

[0070] Set up a reminder mechanism based on the residents’ payment history and behavioral characteristics:

[0071] Among them, Rt represents the push time of the smart reminder, T represents the preset payment date, δ represents the reminder advance time coefficient, and adjusts the advance amount of the reminder.

[0072] Furthermore, payment confirmation and real-time status update are performed in step S3:

[0073] Automatically confirm the arrival of funds according to the contract rules.

[0074] The contract execution results will automatically generate adjustment details and be fed back to residents and property management personnel.

[0075] If the resident fails to follow the adjusted payment plan or meet the litigation standards, a pre-litigation notice will be automatically sent and staff will be notified in real time to intervene.

[0076] Furthermore, in step S3, the payment rules, contract terms and anomaly detection results are converted into smart contracts to automatically execute payment confirmation, overdue penalty calculation and fee adjustment:

[0077] Define the parameters and execution conditions in the smart contract, including automatic processing of payment confirmation, penalty calculation and fee adjustment:

[0078] Payment confirmation automatically checks the payment receipt and updates the payment status.

[0079] The overdue penalty is calculated according to the contract terms.

[0080] Fee adjustment means dynamically adjusting unpaid fees and providing new payment plans;

[0081] Confirm payment to determine whether the resident's actual payment complies with the contract provisions:

[0082] If A i ≥P i Then C i =1, otherwise C i =0, where C i Indicates payment confirmation status, 1 indicates successful payment, 0 indicates unsuccessful payment, A i represents the actual payment amount of household i, P i represents the amount payable by household i, calculated according to the terms of the contract;

[0083] When residents are late for payment, the smart contract automatically calculates the fine amount F i :

[0084] F i =β·Od i ·P i , where F i represents the overdue fine amount of household i, β represents the fine ratio coefficient, Od i Indicates the cumulative number of overdue days for household i since the payment deadline;

[0085] When residents fail to meet payment targets or are overdue, the smart contract dynamically adjusts fees based on the residents’ behavior and generates a new payment plan:

[0086] P′ i =P i +F i -A i , where P′ i represents the new amount payable by household i after adjustment, F i A represents the overdue fine amount for household i, i represents the actual payment amount of household i;

[0087] Smart contracts embed the above logic into an automatic execution process:

[0088] First, check the payment status of the resident and confirm the payment. i =1, the payment status is updated to completed.

[0089] Then, if overdue, enter the overdue penalty calculation and calculate the penalty amount F i ,

[0090] Secondly, according to the payment behavior of the residents, the fee adjustment is performed to calculate the new amount payable P′ i ,

[0091] Finally, the contract automatically generates an adjusted payment plan and notifies residents of the updated payment amount and payment deadline.

[0092] Furthermore, in step S3, the smart contract adjusts the payment cycle and reminder frequency for high-risk residents according to the personalized plan, automatically matching the residents' payment ability and behavioral characteristics:

[0093] Smart contracts need to automatically adjust payment cycles and reminder frequencies based on residents’ payment capabilities, behavioral characteristics, and personalized payment plans. The logic of the contract incorporates the payment behavior characteristics of residents into dynamic adjustments to achieve personalized management of high-risk residents.

[0094] Define parameters and condition settings:

[0095] T i represents the original payment period of household i, which is the standard period stipulated in the contract.

[0096] R i represents the reminder frequency of household i, that is, the number of reminders in a payment cycle,

[0097] c i represents the payment ability score of household i, reflecting its ability to pay the arrears, based on the calculation result in step S2,

[0098] ψ i represents the risk adjustment coefficient of household i, which is calculated based on the household behavior characteristics.

[0099] The smart contract adjusts the payment cycle based on the resident’s ability to pay and the risk adjustment factor:

[0100] Among them, T′ i represents the length of the payment period after adjustment, T irepresents the original payment cycle, λ1 represents the cycle shortening ratio control parameter, e represents the base of the natural logarithm, α represents the sensitivity parameter, which regulates the influence of payment capacity and risk adjustment on cycle adjustment, and β represents the threshold parameter;

[0101] According to the resident’s ability to pay and the risk adjustment factor, the smart contract automatically calculates the adjusted reminder frequency R′ i :

[0102] Among them, R′ i Represents the reminder frequency after adjustment, that is, the number of reminders in each payment cycle after adjustment, R i represents the initial reminder frequency, λ2 represents the reminder frequency growth control parameter, γ represents the sensitivity parameter of frequency adjustment, and θ represents the threshold of frequency adjustment;

[0103] Contract execution process:

[0104] First, the smart contract evaluates the householder’s latest payment capacity c i and the risk adjustment factor ψ i ,

[0105] Then, the new payment period T′ is calculated according to the adjustment formula i and reminder frequency R′ i ,

[0106] Secondly, the contract automatically adjusts the payment schedule, setting new payment deadlines and reminder time nodes.

[0107] Finally, the smart contract pushes the updated plan to the residents and records it.

[0108] The beneficial effects of the present invention are:

[0109] The present invention uses random forests and graph neural networks to monitor and analyze residents' payment behavior in real time. It is no longer limited to static rules and can dynamically identify complex abnormal behaviors. It automatically outputs the risk level of residents through the model, instantly marks high-risk residents, reduces reliance on manual review, and improves the accuracy and timeliness of identification, effectively dealing with residents' long-term or intermittent arrears problems.

[0110] In the present invention, the smart contract automatically executes payment confirmation and overdue penalty calculation, realizes automatic update of payment management, generates dynamically adjusted personalized payment plans according to the residents' payment ability, arrears amount and performance status, and automatically pushes smart reminders before the payment date, reducing manual confirmation steps and speeding up status updates.

[0111] The present invention converts payment rules, contract terms and anomaly detection results into smart contracts, realizes full-process automated management of payment confirmation, overdue fines and fee adjustment tasks, adjusts payment cycles and reminder frequencies based on the latest behaviors of residents, automatically matches residents' payment capabilities and behavioral characteristics, and improves the accuracy of data processing.

[0112] The present invention adopts different management measures for residents of different levels through risk classification, shortens the payment period of high-risk residents, increases reminder frequency, and provides installment options, so that residents can complete payment according to their own payment ability, reduce default rate, and reduce property financial risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0113] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0114] Figure 1 This is a structural diagram of a property payment management system based on big data of the present invention;

[0115] Figure 2 It is a flow chart of the property payment management method based on big data of the present invention. DETAILED DESCRIPTION

[0116] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0117] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0118] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0119] Example 1, reference Figure 1 This embodiment provides a property payment management system based on big data, including:

[0120] The property payment data collection module is responsible for collecting and managing property payment related data, including property contract terms, payment rules, resident payment records and contract performance information;

[0121] The resident labeling and classification module analyzes the payment data of residents, identifies abnormal payment behaviors, outputs abnormal detection results, and classifies the risk level of residents into high, medium, and low;

[0122] The risk assessment module conducts risk assessment and classification of residents based on anomaly detection results and dynamically updates the risk level of residents;

[0123] The high-risk resident payment management module provides personalized payment reminders and management strategies for high-risk residents;

[0124] Smart contract execution and payment automation module converts property contract terms, payment rules and anomaly detection results into smart contracts, and is responsible for all automatically executed operations in the property payment management system, including payment confirmation, overdue processing and fee adjustment;

[0125] The early warning intervention module monitors the payment behavior and feedback of high-risk residents and issues early warnings to residents who may incur financial risks.

[0126] Example 2, reference Figure 2 This embodiment provides a property payment management method based on big data, comprising the following steps:

[0127] Step S1: Payment data collection and abnormal behavior identification.

[0128] Collect the payment data of the residents, integrate and preprocess them, use the random forest model to preliminarily analyze the payment records of the residents, and output the anomaly detection results, that is, the abnormal behavior of the residents.

[0129] Based on abnormal behavior identification, a resident behavior graph is constructed, and the resident's payment mode is combined with the contract performance to output the resident's risk level, which is divided into high, medium and low, and high-risk residents are marked, that is, residents who frequently default or owe large amounts of fees;

[0130] In step S1, abnormal behaviors include frequent overdue payments, large arrears, and intermittent defaults.

[0131] In step S1, the payment data includes property contract terms, payment rules, resident payment records and contract performance information;

[0132] In step S1, the household payment records are analyzed and the abnormality detection results are output as follows:

[0133] Preprocess the collected payment data. Suppose the payment data set is X = {x1, x2, ..., x n}, where each x i Represents a household’s feature vector, including historical payment records, overdue times, and contract performance. The payment behavior features are constructed as the input vector x i =[f1,f2,…,f m ],in,

[0134] Among them, f1 represents the relative overdue frequency of the household, Odds represents the total number of overdue days of the household in a cycle, and Tpp represents the total number of payment days stipulated in the contract.

[0135] Among them, f2 represents the relative arrears amount of the household, Ao represents the total amount of unpaid amount of the household in the period, and Ap represents the amount payable calculated according to the contract terms.

[0136] f3 represents the contract performance score, which is based on the performance of the contract terms and ranges from 0 to 1.

[0137] Building a Random Forest Model Contains T decision trees, and the output of each tree is The model output is: in, represents the predicted anomaly score, which measures the abnormal payment behavior of the resident. Indicates The output of a decision tree is

[0138] Define the threshold τ to identify abnormal behavior, if It is abnormal behavior, otherwise it is normal behavior. τ is set based on the statistical analysis of historical data.

[0139] In step S1, a household behavior graph is constructed based on the anomaly detection results, and the household risk level is output based on the payment mode and contract performance:

[0140] Extract key behavioral features from anomaly detection results, including:

[0141] f freq_delay , represents the frequency of overdue payments of residents in multiple payment cycles,

[0142] f large_debt , indicating the accumulated large amount of arrears of the household,

[0143] f irreg_arrears , indicating the frequency and pattern of episodic arrears of households,

[0144] f contract_compliance represents the household's contract performance score,

[0145] Construct a graph structure, G = (V, E), where V represents the node set, E represents the edge set, and each node v i ∈V represents a household, and its feature vector h is defined i for:

[0146] h i =[f freq_delay ,f large_debt ,f irreg_arrears ,f contract_compliance ], where h i The feature vector representing household i includes overdue frequency, large amount of arrears, intermittent arrears and compliance score;

[0147] Connecting resident nodes with similar payment behaviors, the edge weight is w ij represents the similarity of the behaviors between two residents and is defined as:

[0148] Among them, w ij represents the behavioral similarity weight between residents i and j, |h i -h j | represents the Euclidean distance between eigenvectors, σ represents the smoothing parameter;

[0149] Perform graph convolution on the constructed household behavior graph to obtain the comprehensive behavior characteristics of each household:

[0150] Among them, h′ i represents the feature vector of household i after graph convolution update, represents the set of neighbor nodes of household i, W represents the weight matrix of the graph convolutional layer, and σ represents the activation function;

[0151] The updated feature vector h′ i Input the risk assessment function to obtain the resident’s risk score r i :

[0152] r i =α1f′ freq_delay +α2f′ large_debt +α3f′ irreg_arrears +α4f′ contract_compliance , where r i represents the risk score of household i, α1, α2, α3, α4 represent the characteristic weight coefficients;

[0153] According to the risk score r i , classifying households into high, medium, and low risk levels:

[0154] If r i >θ high , it is classified as high risk level.

[0155] If θ low <r i ≤θ high , it is classified as medium risk level.

[0156] If r i ≤θ low , it is classified as low risk level.

[0157] Among them, θ high represents the high risk threshold, θ low Indicates a low risk threshold;

[0158] Specifically, big data technology and machine learning models are used to conduct real-time analysis of residents' payment behavior, preliminarily identify abnormal behavior through the random forest model, and further use the graph neural network (GNN) to construct a resident behavior graph to quickly and accurately identify abnormal behavior, update the residents' risk level in real time, reduce manual intervention, and improve the accuracy of risk identification.

[0159] Step S2: High-risk resident management,

[0160] Based on the risk level output in step S1, the payment behavior of high-risk residents is tracked, and intelligent reminders and personalized payment plans are pushed to them;

[0161] Smart reminders and personalized payment plans include:

[0162] Before the payment date, payment reminders will be automatically pushed to remind residents to pay on time.

[0163] According to the amount of arrears and contract performance of the residents, the payment ability and behavior characteristics of the residents are analyzed, and a personalized plan, namely the basic payment plan and installment suggestions, is generated and pushed to the residents. The plan is dynamically adjusted based on the amount of arrears and contract performance, reflecting the latest risk status of the residents in real time;

[0164] In step S2, based on the risk level output in step S1, the payment behavior of high-risk residents is tracked, and intelligent reminders and personalized payment plans are pushed to them:

[0165] Identify the high-risk household set from the household risk level output in step S1, set where r i represents the risk score of household i, θ high is a high risk threshold,

[0166] Each high-risk resident Track their payment behavior in real time and build a dynamic feature update model:

[0167] in, represents the behavioral feature vector of household i at time t+1, represents the behavioral feature vector of household i at time t, represents the latest observed behavioral feature vector of household i, including the latest payment behavior and contract performance, and α represents the time decay factor;

[0168] According to the updated feature vector, analyze the behavior trend of the residents and predict their payment ability, and output the payment ability score c i :

[0169] Among them, c i represents the payment ability score of household i, ω1, ω2, ω3, ω4 represent the characteristic weights, Indicates the updated overdue frequency. Indicates the updated large amount of arrears feature. represents the updated intermittent delinquency characteristics, represents the updated contract performance score;

[0170] Based on the ability to pay rating c i , generate a personalized payment plan and set the basic payment amount p i and staging recommendations i :

[0171] p i =γ·Ao, where p i represents the basic payment amount, which is the initial payment amount set by high-risk households based on their actual payment ability; γ represents the payment ratio coefficient, which depends on the household’s risk level and payment ability;

[0172] Among them, s i represents the installment payment amount, which is the remaining arrears divided equally into the next η periods, where η represents the number of installments, and is dynamically adjusted based on the household's risk level and payment ability;

[0173] Set up smart reminder strategies based on residents’ behavioral trends and personalized payment plans:

[0174] Among them, Rt represents the specific time of pushing payment reminders in advance, T represents the payment date, and δ represents the reminder advance parameter, which depends on the payment ability score c i , the lower the score, the earlier the reminder;

[0175] In step S2, based on the amount of the household's arrears and performance, the household's payment ability and behavior characteristics are analyzed to generate a personalized plan:

[0176] For each household i, construct the affordability scoring model c i :

[0177] Among them, c i represents the payment ability score of household i, φ1, φ2, φ3 represent the characteristic weight coefficients, inc i represents the monthly income of household i, Ao i represents the total amount of arrears currently owed by household i, Noop i represents the number of overdue payments of household i in the most recent cycle, r i represents the contract performance score of household i,

[0178] Assess the payment behavior characteristics of households and calculate the risk adjustment coefficient ψ i :

[0179] Among them, ψ i represents the risk adjustment coefficient of household i, κ1, κ2, κ3 represent the characteristic weights, Ap i Tdo is the total amount of fees that household i should pay according to the contract. i represents the cumulative number of overdue days for household i, Cp i represents the length of the payment contract period of household i, f Fd Characteristics indicating that households frequently default over multiple periods;

[0180] Based on the ability to pay scorec i and the risk adjustment factor ψ i , generating the household's basic payment amount p i and staging recommendations i :

[0181] Among them, p i represents the basic payment amount of household i, that is, the first payment amount calculated in a personalized manner, μ represents the basic payment ratio coefficient, and MiniPs represents the minimum payment amount standard;

[0182] Among them, s i represents the installment payment amount of household i, the remaining arrears amount is evenly distributed, ξ represents the installment adjustment factor, and the risk coefficient ψ i Dynamically adjust the number of installments;

[0183] Set up a reminder mechanism based on the residents’ payment history and behavioral characteristics:

[0184] Among them, Rt represents the push time of the smart reminder, T represents the preset payment date, and δ represents the reminder advance time coefficient, which regulates the advance amount of the reminder;

[0185] In terms of contract execution and payment confirmation in the property payment management system, the traditional method still relies on manual operation and manual system recording. The payment standards, overdue processing and fee adjustments in contract management require manual judgment and manual update, which is prone to information inconsistency and untimely processing. Contract management is generally signed by the property company and the residents in paper or electronic documents. The execution of payment, overdue management and adjustment also require manual review and operation. The process is cumbersome and prone to errors. After the residents pay, the system still needs manual confirmation of the arrival of funds, making it difficult to update the residents' payment status in a timely manner.

[0186] The low degree of process automation and insufficient efficiency lead to high financial risks for property management and reduced user experience. In this regard, the present invention optimizes contract management:

[0187] Specifically, for high-risk residents, the system generates personalized payment plans based on anomaly detection results, their payment ability and behavioral characteristics, sets the basic payment amount and installment suggestions, provides residents with more flexible payment options, and dynamically adjusts payment plans based on the residents' real-time payment ability and risk level, which helps reduce the default rate of high-risk residents.

[0188] Step S3: Smart contract execution and automated payment management.

[0189] Convert payment rules, contract terms and anomaly detection results into smart contracts to automatically execute payment confirmation, overdue penalty calculation and fee adjustment.

[0190] The smart contract adjusts the payment cycle and reminder frequency for high-risk residents based on the personalized plan in step S2, automatically matching the residents’ payment ability and behavioral characteristics analyzed in step S2.

[0191] Payment confirmation and real-time status update are performed in step S3:

[0192] Automatically confirm the arrival of funds according to the contract rules.

[0193] The contract execution results will automatically generate adjustment details and be fed back to residents and property management personnel.

[0194] If the resident fails to follow the adjusted payment plan or meet the litigation standards, a pre-litigation notice will be automatically sent and staff will be notified in real time to intervene.

[0195] In step S3, the payment rules, contract terms and anomaly detection results are converted into smart contracts to automatically execute payment confirmation, overdue penalty calculation and fee adjustment:

[0196] Define the parameters and execution conditions in the smart contract, including automatic processing of payment confirmation, penalty calculation and fee adjustment:

[0197] Payment confirmation automatically checks the payment receipt and updates the payment status.

[0198] The overdue penalty is calculated according to the contract terms.

[0199] Fee adjustment means dynamically adjusting unpaid fees and providing new payment plans;

[0200] Confirm payment to determine whether the resident's actual payment complies with the contract provisions:

[0201] If A i ≥P i Then C i =1, otherwise C i =0, where C i Indicates payment confirmation status, 1 indicates successful payment, 0 indicates unsuccessful payment, A i represents the actual payment amount of household i, P i represents the amount payable by household i, calculated according to the terms of the contract;

[0202] When residents are late for payment, the smart contract automatically calculates the fine amount F i :

[0203] F i =β·Od i ·P i , where F i represents the overdue fine amount of household i, β represents the fine ratio coefficient, Od i Indicates the cumulative number of overdue days for household i since the payment deadline;

[0204] When residents fail to meet payment targets or are overdue, the smart contract dynamically adjusts fees based on the residents’ behavior and generates a new payment plan:

[0205] P′ i =P i +F i -A i , where P′ i represents the new amount payable by household i after adjustment, F i A represents the overdue fine amount for household i, i represents the actual payment amount of household i;

[0206] Smart contracts embed the above logic into an automatic execution process:

[0207] First, check the payment status of the resident and confirm the payment. i =1, the payment status is updated to completed.

[0208] Then, if overdue, enter the overdue penalty calculation and calculate the penalty amount Fi ,

[0209] Secondly, according to the payment behavior of the residents, the fee adjustment is performed to calculate the new amount payable P′ i ,

[0210] Finally, the contract automatically generates an adjusted payment plan and notifies residents of the updated payment amount and payment deadline;

[0211] In step S3, the smart contract adjusts the payment cycle and reminder frequency for high-risk residents according to the personalized plan, automatically matching the residents' payment ability and behavioral characteristics:

[0212] Smart contracts need to automatically adjust payment cycles and reminder frequencies based on residents’ payment capabilities, behavioral characteristics, and personalized payment plans. The logic of the contract incorporates the payment behavior characteristics of residents into dynamic adjustments to achieve personalized management of high-risk residents.

[0213] Define parameters and condition settings:

[0214] T i represents the original payment period of household i, which is the standard period stipulated in the contract.

[0215] R i represents the reminder frequency of household i, that is, the number of reminders in a payment cycle,

[0216] c i represents the payment ability score of household i, reflecting its ability to pay the arrears, based on the calculation result in step S2,

[0217] ψ i represents the risk adjustment coefficient of household i, which is calculated based on the household behavior characteristics.

[0218] The smart contract adjusts the payment cycle based on the resident’s ability to pay and the risk adjustment factor:

[0219] Among them, T′ i represents the length of the payment period after adjustment, T i represents the original payment cycle, λ1 represents the cycle shortening ratio control parameter, e represents the base of the natural logarithm, α represents the sensitivity parameter, which regulates the influence of payment capacity and risk adjustment on cycle adjustment, and β represents the threshold parameter;

[0220] According to the resident’s ability to pay and the risk adjustment factor, the smart contract automatically calculates the adjusted reminder frequency R′ i :

[0221] Among them, R′ iRepresents the reminder frequency after adjustment, that is, the number of reminders in each payment cycle after adjustment, R i represents the initial reminder frequency, λ2 represents the reminder frequency growth control parameter, γ represents the sensitivity parameter of frequency adjustment, and θ represents the threshold of frequency adjustment;

[0222] Contract execution process:

[0223] First, the smart contract evaluates the householder’s latest payment capacity c i and the risk adjustment factor ψ i ,

[0224] Then, the new payment period T′ is calculated according to the adjustment formula i and reminder frequency R′ i ,

[0225] Secondly, the contract automatically adjusts the payment schedule, setting new payment deadlines and reminder time nodes.

[0226] Finally, the smart contract pushes the updated plan to the residents and records it.

[0227] Specifically, the payment rules, contract terms and anomaly detection results are converted into smart contracts to achieve automated payment confirmation, overdue penalty calculation and fee adjustment, reduce the time for manual review and processing, and adjust the payment cycle and reminder frequency in real time according to the payment behavior of residents, thereby improving the operating efficiency of the system, increasing the transparency and security of data processing, and reducing potential disputes.

[0228] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A property payment management system based on big data, characterized by: include, The property payment data collection module is responsible for collecting and managing property payment related data, including property contract terms, payment rules, resident payment records and contract performance information; The resident labeling and classification module analyzes the payment data of residents, identifies abnormal payment behaviors, outputs abnormal detection results, and classifies the risk level of residents into high, medium, and low; The risk assessment module conducts risk assessment and classification of residents based on anomaly detection results and dynamically updates the risk level of residents; The high-risk resident payment management module provides personalized payment reminders and management strategies for high-risk residents; Smart contract execution and payment automation module converts property contract terms, payment rules and anomaly detection results into smart contracts, and is responsible for all automatically executed operations in the property payment management system, including payment confirmation, overdue processing and fee adjustment; The early warning intervention module monitors the payment behavior and feedback of high-risk residents and issues early warnings to residents who may incur financial risks.

2. A property payment management method based on big data, characterized in that: Based on the big data-based property payment management system described in claim 1, The following steps are included: Step S1: Payment data collection and abnormal behavior identification. Collect the payment data of the residents, integrate and preprocess them, use the random forest model to preliminarily analyze the payment records of the residents, and output the anomaly detection results, that is, the abnormal behavior of the residents. Based on abnormal behavior identification, a resident behavior graph is constructed, and the resident's payment mode is combined with the contract performance to output the resident's risk level, which is divided into high, medium and low, and high-risk residents are marked, that is, residents who frequently default or owe large amounts of fees; Step S2: High-risk resident management, Based on the risk level output in step S1, the payment behavior of high-risk residents is tracked, and intelligent reminders and personalized payment plans are pushed to them; Smart reminders and personalized payment plans include: Before the payment date, payment reminders will be automatically pushed to remind residents to pay on time. According to the amount of arrears and contract performance of the residents, analyze the residents' payment ability and behavior characteristics, generate personalized plans, namely basic payment plans and installment suggestions, and push them to the residents. The plan is dynamically adjusted based on the amount of arrears and contract performance; Step S3: Smart contract execution and automated payment management. The payment rules, contract terms and anomaly detection results are converted into smart contracts to automatically execute payment confirmation, overdue penalty calculation and fee adjustment. The smart contract adjusts the payment cycle and reminder frequency for high-risk residents based on the personalized plan of step S2, and automatically matches the residents' payment ability and behavioral characteristics analyzed in step S2.

3. A property payment management method based on big data according to claim 2, characterized in that: In step S1, abnormal behaviors include frequent overdue payments, large arrears, and intermittent defaults. In step S1, the payment data includes property contract terms, payment rules, resident payment records and contract performance information.

4. A property payment management method based on big data according to claim 3, characterized in that: In step S1, the household payment records are analyzed and the abnormality detection results are output as follows: Preprocess the collected payment data. Suppose the payment data set is X = {x1, x2, ..., x n }, where each x i Represents a household’s feature vector, including historical payment records, overdue times, and contract performance. The payment behavior features are constructed as the input vector x i =[f1,f2,…,f m ],in, Among them, f1 represents the relative overdue frequency of the household, Odds represents the total number of overdue days of the household in a cycle, and Tpp represents the total number of payment days stipulated in the contract. Among them, f2 represents the relative arrears amount of the household, Ao represents the total amount of unpaid amount of the household in the period, and Ap represents the amount payable calculated according to the contract terms. f3 represents the contract performance score, which is based on the performance of the contract terms and ranges from 0 to 1. Building a Random Forest Model Contains T decision trees, and the output of each tree is The model output is: in, represents the predicted anomaly score, which measures the abnormal payment behavior of the resident. Indicates The output of a decision tree is Define the threshold τ to identify abnormal behavior, if If , it is abnormal behavior, otherwise it is normal behavior. τ is set based on the statistical analysis of historical data.

5. A property payment management method based on big data according to claim 4, characterized in that: In step S1, a household behavior graph is constructed based on the anomaly detection results, and the household risk level is output based on the payment mode and contract performance: Extract key behavioral features from anomaly detection results, including: f freq_delay , represents the frequency of overdue payments of residents in multiple payment cycles, f large_debt , indicating the accumulated large amount of arrears of the household, f irreg_arrears , indicating the frequency and pattern of episodic arrears of households, f contract_compliance represents the household's contract performance score, Construct a graph structure, G = (V, E), where V represents the node set, E represents the edge set, and each node v i ∈V represents a household, and its feature vector h is defined i for: h i =[f freq_delay ,f large_debt ,f irreg_arrears ,f contract_compliance ], where h i The feature vector representing household i includes overdue frequency, large amount of arrears, intermittent arrears and compliance score; Connecting resident nodes with similar payment behaviors, the edge weight is w ij represents the similarity of the behaviors between two residents and is defined as: Among them, w ij represents the behavioral similarity weight between residents i and j, |h i -h j | represents the Euclidean distance between eigenvectors, σ represents the smoothing parameter; Perform graph convolution on the constructed household behavior graph to obtain the comprehensive behavior characteristics of each household: Among them, h′ i represents the feature vector of household i after graph convolution update, represents the set of neighbor nodes of household i, W represents the weight matrix of the graph convolutional layer, and σ represents the activation function; The updated feature vector h′ i Input the risk assessment function to obtain the resident’s risk score r i : r i =α1f′ freq_delay +α2f′ large_debt +α3f′ irreg_arrears +α4f′ contract_compliance , where r i represents the risk score of household i, α1, α2, α3, α4 represent the characteristic weight coefficients; According to the risk score r i , classifying households into high, medium, and low risk levels: If r i >θ high , it is classified as high risk level. If θ low <r i ≤θ high , it is classified as medium risk level. If r i ≤θ low , it is classified as low risk level. Among them, θ high represents the high risk threshold, θ low Indicates a low risk threshold.

6. A property payment management method based on big data according to claim 5, characterized in that: In step S2, based on the risk level output in step S1, the payment behavior of high-risk residents is tracked, and intelligent reminders and personalized payment plans are pushed to them: Identify the high-risk household set from the household risk level output in step S1, set where r i represents the risk score of household i, θ high is a high risk threshold, Each high-risk resident Track their payment behavior in real time and build a dynamic feature update model: in, represents the behavioral feature vector of household i at time t+1, represents the behavioral feature vector of household i at time t, represents the latest observed behavioral feature vector of household i, including the latest payment behavior and contract performance, and α represents the time decay factor; According to the updated feature vector, analyze the behavior trend of the residents and predict their payment ability, and output the payment ability score c i : Among them, c i represents the payment ability score of household i, ω1, ω2, ω3, ω4 represent the characteristic weights, Indicates the updated overdue frequency. Indicates the updated large amount of arrears feature. represents the updated intermittent delinquency characteristics, represents the updated contract performance score; Based on the ability to pay rating c i , generate a personalized payment plan and set the basic payment amount p i and staging recommendations i : p i =γ·Ao, where p i represents the basic payment amount, which is the initial payment amount set by high-risk households based on their actual payment ability; γ represents the payment ratio coefficient, which depends on the household’s risk level and payment ability; Among them, s i represents the installment payment amount, which is the remaining arrears divided equally into the next η periods, where η represents the number of installments, and is dynamically adjusted based on the household's risk level and payment ability; Set up smart reminder strategies based on residents’ behavioral trends and personalized payment plans: Among them, Rt represents the specific time of pushing payment reminders in advance, T represents the payment date, and δ represents the reminder advance parameter, which depends on the payment ability score c i , the lower the score, the earlier the reminder.

7. A property payment management method based on big data according to claim 6, characterized in that: In step S2, based on the amount of the household's arrears and performance, the household's payment ability and behavior characteristics are analyzed to generate a personalized plan: For each household i, construct the affordability scoring model c i : Among them, c i represents the payment ability score of household i, φ1, φ2, φ3 represent the characteristic weight coefficients, inc i represents the monthly income of household i, Ao i represents the total amount of arrears currently owed by household i, Noop i represents the number of overdue payments of household i in the most recent cycle, r i represents the contract performance score of household i, Assess the payment behavior characteristics of households and calculate the risk adjustment coefficient ψ i : Among them, ψ i represents the risk adjustment coefficient of household i, κ1, κ2, κ3 represent the characteristic weights, Ap i Tdo is the total amount of fees that household i should pay according to the contract. i represents the cumulative number of overdue days for household i, Cp i represents the length of the payment contract period of household i, f Fd Characteristics indicating that households frequently default over multiple periods; Based on the ability to pay scorec i and the risk adjustment factor ψ i , generating the household's basic payment amount p i and staging recommendations i : Among them, p i represents the basic payment amount of household i, that is, the first payment amount calculated in a personalized manner, μ represents the basic payment ratio coefficient, and MiniPs represents the minimum payment amount standard; Among them, s i represents the installment payment amount of household i, the remaining arrears amount is evenly distributed, ξ represents the installment adjustment factor, and the risk coefficient ψ i Dynamically adjust the number of installments; Set up a reminder mechanism based on the residents’ payment history and behavioral characteristics: Among them, Rt represents the push time of the smart reminder, T represents the preset payment date, δ represents the reminder advance time coefficient, and adjusts the advance amount of the reminder.

8. A property payment management method based on big data according to claim 7, characterized in that: Payment confirmation and real-time status update are performed in step S3: Automatically confirm the arrival of funds according to the contract rules. The contract execution results will automatically generate adjustment details and be fed back to residents and property management personnel. If the resident fails to follow the adjusted payment plan or meet the litigation standards, a pre-litigation notice will be automatically sent and staff will be notified in real time to intervene.

9. A property payment management method based on big data according to claim 8, characterized in that: In step S3, the payment rules, contract terms and anomaly detection results are converted into smart contracts to automatically execute payment confirmation, overdue penalty calculation and fee adjustment: Define the parameters and execution conditions in the smart contract, including automatic processing of payment confirmation, penalty calculation and fee adjustment: Payment confirmation automatically checks the payment receipt and updates the payment status. The overdue penalty is calculated according to the contract terms. Fee adjustment means dynamically adjusting unpaid fees and providing new payment plans; Confirm payment to determine whether the resident's actual payment complies with the contract provisions: If A i ≥P i Then C i =1, otherwise C i =0, where C i Indicates payment confirmation status, 1 indicates successful payment, 0 indicates unsuccessful payment, A i represents the actual payment amount of household i, P i represents the amount payable by household i, calculated according to the terms of the contract; When residents are late for payment, the smart contract automatically calculates the fine amount F i : F i =β·Od i ·P i , where F i represents the overdue fine amount of household i, β represents the fine ratio coefficient, Od i Indicates the cumulative number of overdue days for household i since the payment deadline; When residents fail to meet payment targets or are overdue, the smart contract dynamically adjusts fees based on the residents’ behavior and generates a new payment plan: P i ′=P i +F i -A i , where P i ′ represents the new amount payable by household i after adjustment, F i A represents the overdue fine amount for household i, i represents the actual payment amount of household i.

10. A property payment management method based on big data according to claim 9, characterized in that: In step S3, the smart contract adjusts the payment cycle and reminder frequency for high-risk residents according to the personalized plan, automatically matching the residents' payment ability and behavioral characteristics: Define parameters and condition settings: T i represents the original payment period of household i, which is the standard period stipulated in the contract. R i represents the reminder frequency of household i, that is, the number of reminders in a payment cycle, c i represents the payment ability score of household i, reflecting its ability to pay the arrears, based on the calculation result in step S2, ψ i represents the risk adjustment coefficient of household i, which is calculated based on the household behavior characteristics. The smart contract adjusts the payment cycle based on the resident’s ability to pay and the risk adjustment factor: Among them, T i ′ represents the length of the payment period after adjustment, T i represents the original payment cycle, λ1 represents the cycle shortening ratio control parameter, e represents the base of the natural logarithm, α represents the sensitivity parameter, which regulates the influence of payment capacity and risk adjustment on cycle adjustment, and β represents the threshold parameter; According to the resident’s ability to pay and the risk adjustment factor, the smart contract automatically calculates the adjusted reminder frequency R′ i : Among them, R′ i Represents the reminder frequency after adjustment, that is, the number of reminders in each payment cycle after adjustment, R i represents the initial reminder frequency, λ2 represents the reminder frequency growth control parameter, γ represents the sensitivity parameter of frequency adjustment, and θ represents the threshold of frequency adjustment.

Images