Safety lease crowd evaluation system of product lease platform

Abstract

The invention relates to the technical field of safe lease crowd assessment, is used for solving the problems that a traditional product lease platform lacks multi-dimensional data monitoring in the aspect of user safety assessment and risk user groups are inaccurately identified, and particularly discloses a safe lease crowd assessment system of a product lease platform. Comprising a data acquisition module, a server, a data analysis module, a risk assessment module, a delimiting analysis module, a verification analysis module and a display terminal. According to the method, the corresponding risk values are determined by monitoring the basic data parameters, the external data parameters and the behavior data parameters of the user, the lease risk assessment model is constructed based on the risk values, the comprehensive risk assessment value is calculated to determine the risk type state, and the lease risk assessment model is constructed according to the user characteristics corresponding to the risk type state. The features of different risk groups are accurately identified, and the target feature information interval is determined accordingly, so that accurate classification of user groups is realized, lease risk assessment accuracy is improved, and lease transaction safety is effectively guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of safe rental population assessment, and in particular to a safe rental population assessment system for a product rental platform. Background Art

[0002] With the rapid development of the internet, product rental platforms have gradually become an important channel for more and more consumers to conduct online rental transactions. However, the special characteristics of online transactions make user security and reliability always the core issues in the platform's safe operation. During the rental process, one of the biggest challenges facing the platform is how to accurately identify and screen qualified and high-quality rental user groups to effectively prevent potential losses caused by dishonest and high-risk users.

[0003] Currently, many product leasing platforms still have some shortcomings in security assessment. Traditional assessment methods often rely on preset assessment conditions and are too simplistic. This rigid assessment model can be easily circumvented by some users through falsified information or other improper means, thereby reducing the accuracy and effectiveness of the assessment. This makes it difficult for platforms to effectively identify high-risk users in their leasing business, thereby increasing the uncertainty and risk of platform operations. In addition, traditional assessment systems lack the ability to dynamically adjust based on the characteristics of risky users, making it impossible to timely and accurately define high-risk user groups. This static assessment method is not only unable to adapt to the complex and changing network environment, but also fails to meet the platform's demand for refined risk management.

[0004] In order to solve the above-mentioned defects, a technical solution is now provided. Summary of the Invention

[0005] The purpose of this invention is to solve the limitations of traditional product rental platforms in security assessment, such as the inability to conduct multi-dimensional data monitoring and analysis of users, comprehensively assess users' risk levels, and accurately identify and dynamically adjust the delineation of high-risk user groups. Instead, a security rental population assessment system for product rental platforms is proposed.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A safe rental crowd assessment system for a product rental platform, comprising:

[0008] A data analysis module is configured to monitor the user's basic data parameters through a first data analysis unit to obtain a rental category value, a rental value, and a rental time value to determine a first risk value; monitor the user's external data parameters through a second data analysis unit to obtain a credit score fluctuation value, a credit impact value, and a credit support value to determine a second risk value; and monitor the user's behavioral data parameters through a third data analysis unit to obtain a usage loss rating index and a usage return rating index to determine a third risk value;

[0009] The risk assessment module is used to extract the user's first risk value, second risk value and third risk value and substitute them into the rental risk assessment model to obtain a comprehensive risk assessment value to determine the risk type status;

[0010] A delineation analysis module is used to analyze and process the delineation results of the user according to the risk type status to determine the target feature information interval, which includes a high-risk feature information interval and a low-risk feature information interval;

[0011] Determine users who meet the high-risk characteristic information range as high-risk users;

[0012] Determine users who meet the low-risk characteristic information range as low-risk users;

[0013] The verification and analysis module is used to perform verification and analysis processing on the delineation results and obtain a verification pass signal or a verification fail signal.

[0014] Furthermore, the basic data parameters of the user are monitored. The specific process is as follows:

[0015] Real-time acquisition of the user's rental records within a unit time period, extraction of the categories of each rental product from the rental records, matching and analyzing the categories of each rental product with a pre-set product category table to determine the category level of each rental product, and assigning a corresponding numerical value in sequence according to the category level. This results in the category value of each household's rental product, and the average of the category values of each rental product is calculated to obtain the user's rental product category value Fp;

[0016] Extract the price of each rental product from the rental record, and calculate the average of the prices of each rental product to obtain the user's rental value Fj;

[0017] The rental duration of each rental product is extracted from the rental record, and the rental duration of each rental product is averaged to obtain the user's rental duration value Fs.

[0018] Furthermore, the first risk value is determined, and the specific process is as follows:

[0019] Extract the user's rental category value Fp, rental value Fj and rental time value Fs and normalize them according to the formula: Fpjszg =η1×Fp 2 +η2×ln(Fj)+η3×Fs 2 , calculate the user's first risk value F pjszg , where η1, η2, and η3 represent the set weight coefficients respectively, and η1>η2>η3.

[0020] Furthermore, the user's external data parameters are monitored. The specific process is as follows:

[0021] Obtain the user's credit record within a unit time period in real time, extract the initial credit score and the credit score at each monitoring time point from the credit record, calculate the difference between the credit score at each monitoring time point and the initial credit score and take the absolute value to obtain the credit score difference at each monitoring time point to determine the user's credit score fluctuation value Xp;

[0022] At the same time, the credit score difference at each monitoring time point is compared and analyzed with the set credit score difference threshold. If the credit score difference at a monitoring time point is greater than the set credit score difference threshold, the monitoring time point is marked as an abnormal point;

[0023] Obtain the loan records corresponding to each anomaly, and extract the number of loan platforms, overdue amount, and overdue duration from the loan records to determine the user's credit impact value Xd;

[0024] The replacement payment frequency, payment failure ratio and average payment amount are extracted from the credit record to determine the user's credit rating value Xz.

[0025] Furthermore, the second risk value is determined, and the specific process is as follows:

[0026] Extract the user's credit score fluctuation value Xp, credit impact value Xd and credit rating value Xz and normalize them according to the formula: zfpgz =(Xp+1)×η4+(Xd+2)×η5+(Xz+3)×η6, calculate the user's second risk value X zfpgz , where η4, η5 and η6 represent the set weight coefficients respectively, and η4>η5>η6.

[0027] Furthermore, the user's behavior data parameters are monitored. The specific process is as follows:

[0028] Obtain the user's usage records within a unit time period in real time, and extract the maximum loss rate, minimum loss rate, and median loss rate of each rental product from the usage records to determine the user's usage loss evaluation index Sh. The median loss rate is specifically obtained by arranging the loss rates of each rental product monitored in the usage records in descending order to obtain a loss rate sequence of the rental products, and extracting the loss rate corresponding to the median from the loss rate sequence as the median loss rate;

[0029] The return rate and return time of each rental product are extracted from the usage records to determine the user's usage return evaluation index Sg.

[0030] Furthermore, the third risk value is determined, and the specific process is as follows:

[0031] Extract the user's usage damage evaluation index Sh and usage return evaluation index Sg and perform normalization processing according to the formula: Calculate the user's third risk value S sypin , where u represents a natural constant, η7 and η8 represent set weight coefficients, and η7>η8.

[0032] Furthermore, the risk type status is determined. The specific process is as follows:

[0033] Extract the user's first risk value F pjszg , the second risk value X zfpgz and the third risk value S sypin Substitute into the leasing risk assessment model, according to the risk assessment formula: Calculate the user's comprehensive risk assessment value FXS 综合 , where β1, β2 and β3 represent the correction coefficients of the first risk value, the second risk value and the third risk value respectively;

[0034] Compare and analyze the user's comprehensive risk assessment value with the preset reference risk interval. When the user's comprehensive risk assessment value is greater than the maximum value of the reference risk interval, the user is determined to be in a Class I risk state. When the user's comprehensive risk assessment value is within the reference risk interval, the user is determined to be in a Class II risk state. When the user's comprehensive risk assessment value is less than the minimum value of the reference risk interval, the user is determined to be in a Class III risk state.

[0035] The risk type status is composed of Class I risk status, Class II risk status and Class III risk status.

[0036] Furthermore, the target feature information interval is determined, and the specific process is as follows:

[0037] Count the users who are judged to be in a certain risk state within a period of time to obtain a set of high-risk users;

[0038] Count the users who are judged as Class II or Class III risk status within a period of time to obtain a set of low-risk users;

[0039] The characteristic information of the corresponding users is extracted from the high-risk user set and the low-risk user set respectively. Multiple intervals are set for each characteristic information, and the characteristic information of each user is matched with the corresponding interval. Then, statistics are performed on each characteristic information interval, and the number of users in each interval is calculated. The characteristic information interval with the largest number of users is identified as the target characteristic information interval.

[0040] Furthermore, the results are verified and analyzed. The specific process is as follows:

[0041] By obtaining the damage rate, overdue rate and revenue within a period of time after the demarcation results, the verification value is determined and compared with the set reference verification range for analysis;

[0042] If the verification value is within the set reference verification range, a verification pass signal is generated;

[0043] If the verification value is outside the set reference verification range, a verification failure signal is generated and the user's re-demarcation is triggered. During the re-demarcation process, the user's demarcation results will be adjusted according to the new analysis results, and verification analysis will be performed again until a verification pass signal is generated.

[0044] Compared with the prior art, the technical solution provided by the present invention has the following beneficial effects:

[0045] 1. This invention comprehensively monitors and analyzes users' basic data, external data, and behavioral data to comprehensively assess their risk levels, calculate corresponding risk values, and construct a risk assessment model based on these values. This multi-dimensional analysis method enables a more accurate distinction between high-risk and low-risk users, reduces the impact of forged information or circumvention methods on assessment results, and thus significantly improves the security of the platform and reduces the risk of misjudgment and missed judgments.

[0046] 2. The present invention adopts a dynamic risk assessment mechanism to calculate risk values based on real-time user data and dynamically adjust risk classification based on user characteristics (such as age, occupation, and education). This flexible mechanism can quickly adapt to changing network environments, avoid lag effects, and ensure that assessment results are always accurate and effective.

[0047] 3. The present invention, through in-depth analysis of user characteristics (such as age, occupation, education level, etc.), accurately identifies the characteristics of different risk groups, and then implements targeted risk control measures for different groups, thereby achieving fine division of users and personalized risk control management, which not only improves the operational efficiency of the platform, but also enhances customer satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0049] Figure 1 It is the overall module block diagram of the present invention. DETAILED DESCRIPTION

[0050] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0051] like Figure 1 As shown, a safe rental population assessment system for a product rental platform includes: a data acquisition module, a server, a data analysis module, a risk assessment module, a demarcation analysis module, a verification analysis module, and a display terminal, wherein the data analysis module includes a first data analysis unit, a second data analysis unit, and a third data analysis unit;

[0052] The data collection module is used to collect the user's rental status parameters and verification status parameters, and send them to the data analysis module and verification analysis module respectively through the server;

[0053] It should be noted that the user data collected comes from users who rent on the rental platform, and the data parameters used include not only the data collected by the rental platform itself, but also relevant data from third-party service platforms. In addition, the acquisition and use of all data strictly comply with relevant laws and regulations on privacy protection to ensure compliance with privacy requirements and take necessary security measures to protect users' personal information;

[0054] The data analysis module is used to receive the user's rental status parameters, wherein the rental status parameters include basic data parameters, external data parameters and behavioral data parameters;

[0055] Preferably, the basic data parameters of the user are monitored by the first data analysis unit, thereby analyzing and processing the first risk value of the user. The specific operation process is as follows:

[0056] Real-time acquisition of the user's rental records within a unit time period. Extract the categories of each rental product from the rental records. Match and analyze the categories of each rental product with the pre-set product category table to determine the category level of each rental product. Each category has a corresponding category level, and a corresponding numerical value is assigned in sequence according to the category level. This yields the category value of each household's rental product. The category values of each rental product are averaged to obtain the user's rental product category value, recorded as Fp.

[0057] Extract the price of each rental product from the rental record, and calculate the average of the prices of each rental product to obtain the user's rental value, which is recorded as Fj;

[0058] Extract the rental duration of each rental product from the rental record, calculate the average of the rental duration of each rental product, and obtain the user's rental duration value, recorded as Fs;

[0059] Extract the user's rental category value Fp, rental value Fj and rental time value Fs and normalize them according to the formula: F pjszg =η1×Fp 2 +η2×ln(Fj)+η3×Fs 2 , calculate the user's first risk value F pjszg , where η1, η2, and η3 represent the set weight coefficients respectively, the sum of the weight coefficients is 1 (i.e., η1+η2+η3=1), and η1>η2>η3;

[0060] Preferably, the user's external data parameters are monitored by the second data analysis unit, thereby analyzing and processing the user's second risk value. The specific operation process is as follows:

[0061] Obtain the user's credit record within a unit time period in real time, extract the initial credit score from the credit record, that is, the user's credit score when first applying for a lease, and the credit score at each monitoring time point, calculate the difference between the credit score at each monitoring time point and the initial credit score and take the absolute value to obtain the credit score difference at each monitoring time point, denoted as pfz i , according to the formula: Calculate the user's credit score fluctuation value Xp, where e represents a natural constant, i represents the number of each monitoring time point, and i = 1, 2, 3...n1, n1 represents the total number of monitoring time point numbers, pfz i-1 represents the credit score difference at the i-1th monitoring time point, Δpfz1 represents the reference difference of the credit score difference between each monitoring time point, and pfz * represents the reference credit score difference, Δpfz2 represents the reference difference between the credit score difference at each monitoring time point and the reference credit score difference, a1 and a2 represent the set weight factors respectively;

[0062] At the same time, the credit score difference at each monitoring time point is compared and analyzed with the set credit score difference threshold. If the credit score difference at a monitoring time point is greater than the set credit score difference threshold, the monitoring time point is marked as an abnormal point;

[0063] The loan records corresponding to each abnormal point are obtained, and the number of loan platforms, overdue amount and overdue duration are extracted from the loan records, and marked as dpt respectively. j 、dje j and DSC j , j represents the number of each abnormal point, and j = 1, 2, 3...n2, n2 represents the total number of abnormal point numbers, according to the formula: Calculate the user's credit impact value Xd, where b1, b2, and b3 represent the set weight factors respectively;

[0064] Extract the replacement payment frequency, payment failure ratio, and average payment amount from the credit records and mark them as zfp, zsc, and zje respectively, according to the formula: Calculate the user's credit rating Xz, where b4, b5, and b6 represent the set weight factors respectively;

[0065] It should be noted that the solution to the payment change frequency is to count the number of times a user changes their payment method within a unit time period and calculate the payment change frequency according to the following formula: The higher the replacement frequency, the more unstable the user's payment behavior is;

[0066] The solution to the payment failure ratio is to count the number of payment failures and the total number of payments made by users within a unit time period and calculate the payment failure ratio according to the following formula: A higher payment failure rate indicates that the user's payment status is unstable;

[0067] The average payment amount is calculated by extracting the amount paid by the user each time within a unit time period and calculating the average payment amount according to the calculation formula: The higher the average payment amount, the higher the user's consumption level;

[0068] Extract the user's credit score fluctuation value Xp, credit impact value Xd and credit rating value Xz and normalize them according to the formula: zfpgz =(Xp+1)×η4+(Xd+2)×η5+(Xz+3)×η6, calculate the user's second risk value X zfpgz , where η4, η5 and η6 represent the set weight coefficients respectively, and η4>η5>η6;

[0069] Preferably, the user's behavior data parameters are monitored by the third data analysis unit, thereby analyzing and processing the user's third risk value. The specific operation process is as follows:

[0070] Obtain the user's usage records within a unit time period in real time, and extract the maximum loss rate, minimum loss rate and median loss rate of each rental product from the usage records, and mark them as SH max f SH min f and SH mzd f , where the median loss rate is specifically: arrange the loss rates of each rental product monitored in the usage record in descending order to obtain a loss rate sequence of the rental products, and extract the loss rate corresponding to the median from the loss rate sequence as the median loss rate; according to the formula: Calculate the user's usage damage index Sh, where f represents the number of each rental product, and f = 1, 2, 3...m, m represents the total number of rental product numbers, SH 参考 It represents the set reference loss rate, ΔSH represents the allowable loss difference, c1, c2, and c3 represent the weight factors corresponding to the maximum loss rate difference, the minimum loss rate difference, and the median loss rate difference, respectively;

[0071] It should be noted that the loss rate of a leased product refers to the degree of physical wear or damage to the product during the lease. Specifically, the loss rate can be calculated by the ratio of the use loss of the leased product to its original value. For example, if a leased electronic device loses 10% of its functions or components after a certain period of use, its loss rate is 10%. More specifically, if the original value of a leased electronic device is 1,000 yuan and it loses 100 yuan worth of it during use, then the loss rate is 10%.

[0072] Extract the return rate and return time of each rental product from the usage record and mark them as GH hld f and GH hsz f , according to the formula: Calculate the user's usage rating index Sg, where GH hld * and GH hsz * They represent the set reference return rate and reference return time respectively, c4 and c5 represent the weight factors corresponding to the return rate difference and return time difference respectively;

[0073] Extract the user's usage damage evaluation index Sh and usage return evaluation index Sg and perform normalization processing according to the formula: Calculate the user's third risk value S sypin , where u represents a natural constant, η7 and η8 represent set weight coefficients, and η7>η8;

[0074] The risk assessment module is used to evaluate and analyze the user's rental risk status. The specific process is as follows:

[0075] Extract the user's first risk value F pjszg , the second risk value X zfpgz and the third risk value S sypin Substitute into the leasing risk assessment model, according to the risk assessment formula: Calculate the user's comprehensive risk assessment value FXS 综合 , where β1, β2 and β3 represent the correction coefficients of the first risk value, the second risk value and the third risk value respectively;

[0076] Compare and analyze the user's comprehensive risk assessment value with the preset reference risk range, and classify the user according to the following rules:

[0077] Class I risk status: When the user's comprehensive risk assessment value is greater than the maximum value of the reference risk range, the user is judged to be in Class I risk status;

[0078] Category II risk status: When the user's comprehensive risk assessment value is within the reference risk range, the user is judged to be in Category II risk status;

[0079] Category III risk status: When the user's comprehensive risk assessment value is less than the minimum value of the reference risk range, the user is judged to be in Category III risk status;

[0080] The risk type status is composed of the first risk status, the second risk status and the third risk status;

[0081] The delineation analysis module is used to analyze and process the delineation results of users based on their risk type status. The specific operation process is as follows:

[0082] Count the users who are judged to be in a certain risk state within a period of time to obtain a set of high-risk users;

[0083] Count the users who are judged as Class II or Class III risk status within a period of time to obtain a set of low-risk users;

[0084] Extracting characteristic information of corresponding users from the high-risk user set and the low-risk user set, respectively. The characteristic information includes but is not limited to: age, occupation, and education level;

[0085] Set multiple intervals for each feature information, match each user's feature information with the corresponding interval, then perform statistics on each feature information interval, calculate the number of users in each interval, identify the feature information interval with the largest number of users, and use it as the target feature information interval. The target feature information interval includes a high-risk feature information interval and a low-risk feature information interval.

[0086] In a specific implementation example, taking age in feature information as an example, specifically: extract the age of each user from the high-risk user set, and set multiple age intervals, for example, age: 18-25 years old, 26-35 years old, 36-45 years old, and over 46 years old;

[0087] Match each user's age to the corresponding interval. For example, if a user is 30 years old, the age matching interval is: 26-35 years old;

[0088] Count the number of users in each age range. If the 26-35 age range has the largest number of users, then this age range is considered a high-risk age range.

[0089] Similarly, taking age in feature information as an example, specifically: extract the age of each user from the low-risk user set, and set multiple age intervals, for example, age: 18-25 years old, 26-35 years old, 36-45 years old, and over 46 years old;

[0090] Match each user's age to the corresponding interval. For example, if a user is 20 years old, the age matching interval is: 18-25 years old;

[0091] Count the number of users in each age range. If the 18-25 age range has the largest number of users, then this age range is considered a low-risk age range.

[0092] High-risk feature information ranges are used as criteria for screening high-risk users. Users who meet this range will be identified as high-risk users. The platform will conduct enhanced verification and strict risk control measures on these users, and display notifications on the display terminal;

[0093] The low-risk feature information range is used as the standard for screening low-risk users. Users who meet this range are judged as low-risk users. The platform can push more product rental information to these users to promote their rental behavior and display notifications on the display terminal;

[0094] In a specific implementation example, suppose a rental platform is providing equipment rental services. By analyzing the risk status of different user groups, the following trends are found:

[0095] High-risk users: Most of them are freelancers aged 26-35 with a bachelor's degree or below. The platform will strengthen credit review for these users before leasing equipment, and limit the types and amounts of equipment they can rent.

[0096] Low-risk users: Most of them are corporate employees aged 36-45, with a master's degree or above. The platform pushes more high-value, high-quality equipment rental information to this type of users, and provides exclusive rental discounts;

[0097] The verification analysis module is used to receive the user's verification status parameters and perform verification analysis and processing on the demarcation results. The specific operation process is as follows:

[0098] By obtaining the damage rate, overdue rate and revenue within a period of time after the demarcation result, and marking them as YZ1, YZ2 and YZ3 respectively, the three values are extracted and normalized according to the formula: Calculate the verification value YZP, where λ1, λ2, and λ3 represent the set weight coefficients, and λ1>λ2>λ3;

[0099] It should be noted that the damage rate refers to the total damage ratio of all rental products on the product rental platform within a period of time after the implementation of the delineation results. Specifically, the damage rate is determined by calculating the ratio of the actual damage value of the rental products on the platform during the period of time to the total value of the rental products. The overdue rate refers to the overdue return ratio of all rental products on the platform within a period of time after the implementation of the delineation results. The overdue rate is calculated by counting the ratio of the number of rental products that are not returned on time to the total number of rental products. The revenue refers to the total income obtained by the platform from rental products within a period of time after the implementation of the delineation results. This income includes the amount of all rental transactions and reflects the economic benefits of the platform during the period of time.

[0100] Compare and analyze the validation value with the set reference validation interval;

[0101] If the verification value is within the set reference verification range, a verification pass signal is generated and displayed on the display terminal for notification;

[0102] If the verification value is outside the set reference verification range, a verification failure signal is generated and the user's re-demarcation is triggered. During the re-demarcation process, the user's demarcation results will be adjusted according to the new analysis results, and verification analysis will be performed again until a verification pass signal is generated and a notification will be displayed on the display terminal.

[0103] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A safe rental crowd assessment system for a product rental platform, comprising: The data analysis module is configured to monitor the basic data parameters of the user through a first data analysis unit to determine a first risk value, and is characterized in that: The user's external data parameters are monitored by the second data analysis unit to obtain a credit score fluctuation value, a credit impact value, and a credit support value to determine a second risk value. The user's behavioral data parameters are monitored by the third data analysis unit to obtain a usage loss rating index and a usage return rating index to determine a third risk value. The risk assessment module is used to extract the user's first risk value, second risk value and third risk value and substitute them into the rental risk assessment model to obtain a comprehensive risk assessment value to determine the risk type status; A delineation analysis module is used to analyze and process the delineation results of the user according to the risk type status to determine the target feature information interval, which includes a high-risk feature information interval and a low-risk feature information interval; Determine users who meet the high-risk characteristic information range as high-risk users; Users who meet the low-risk characteristic information range are judged as low-risk users.

2. The safe rental crowd evaluation system for a product rental platform according to claim 1 is characterized in that: Monitor the user's basic data parameters. The specific process is as follows: Real-time acquisition of the user's rental records within a unit time period, extraction of the categories of each rental product from the rental records, matching and analyzing the categories of each rental product with a pre-set product category table to determine the category level of each rental product, and assigning a corresponding numerical value in sequence according to the category level. This results in the category value of each household's rental product, and the average of the category values of each rental product is calculated to obtain the user's rental product category value Fp; Extract the price of each rental product from the rental record, and calculate the average of the prices of each rental product to obtain the user's rental value Fj; The rental duration of each rental product is extracted from the rental record, and the rental duration of each rental product is averaged to obtain the user's rental duration value Fs.

3. The safe rental crowd evaluation system for a product rental platform according to claim 1 is characterized in that: Determine the first risk value. The specific process is as follows: Extract the user's rental category value Fp, rental value Fj and rental time value Fs and normalize them according to the formula: F pjszg =η1×Fp 2 +η2×ln(Fj)+η3×Fs 2 , calculate the user's first risk value F pjszg , where η1, η2, and η3 represent the set weight coefficients respectively, and η1>η2>η3.

4. The safe rental crowd evaluation system for a product rental platform according to claim 1 is characterized in that: Monitor the user's external data parameters. The specific process is as follows: Obtain the user's credit record within a unit time period in real time, extract the initial credit score and the credit score at each monitoring time point from the credit record, calculate the difference between the credit score at each monitoring time point and the initial credit score and take the absolute value to obtain the credit score difference at each monitoring time point to determine the user's credit score fluctuation value Xp; At the same time, the credit score difference at each monitoring time point is compared and analyzed with the set credit score difference threshold. If the credit score difference at a monitoring time point is greater than the set credit score difference threshold, the monitoring time point is marked as an abnormal point; Obtain the loan records corresponding to each anomaly, and extract the number of loan platforms, overdue amount, and overdue duration from the loan records to determine the user's credit impact value Xd; The replacement payment frequency, payment failure ratio and average payment amount are extracted from the credit record to determine the user's credit rating value Xz.

5. The safe rental crowd evaluation system for a product rental platform according to claim 1 is characterized in that: Determine the second risk value. The specific process is as follows: Extract the user's credit score fluctuation value Xp, credit impact value Xd and credit rating value Xz and normalize them according to the formula: zfpgz =(Xp+1)×η4+(Xd+2)×η5+(Xz+3)×η6, calculate the user's second risk value X zfpgz , where η4, η5 and η6 represent the set weight coefficients respectively, and η4>η5>η6.

6. The safe rental crowd evaluation system for a product rental platform according to claim 1 is characterized in that: Monitor the user's behavioral data parameters. The specific process is as follows: Obtain the user's usage records within a unit time period in real time, and extract the maximum loss rate, minimum loss rate, and median loss rate of each rental product from the usage records to determine the user's usage loss evaluation index Sh. The median loss rate is specifically obtained by arranging the loss rates of each rental product monitored in the usage records in descending order to obtain a loss rate sequence of the rental products, and extracting the loss rate corresponding to the median from the loss rate sequence as the median loss rate; The return rate and return time of each rental product are extracted from the usage records to determine the user's usage return evaluation index Sg.

7. The safe rental crowd evaluation system for a product rental platform according to claim 1 is characterized in that: Determine the third risk value. The specific process is as follows: Extract the user's usage damage evaluation index Sh and usage return evaluation index Sg and perform normalization processing according to the formula: Calculate the user's third risk value S sypin , where u represents a natural constant, η7 and η8 represent set weight coefficients, and η7>η8.

8. The safe rental crowd evaluation system for a product rental platform according to claim 1 is characterized in that: Determine the risk type status. The specific process is as follows: Extract the user's first risk value F pjszg , the second risk value X zfpgz and the third risk value S sypin Substitute into the leasing risk assessment model, according to the risk assessment formula: Calculate the user's comprehensive risk assessment value FXS 综合 , where β1, β2 and β3 represent the correction coefficients of the first risk value, the second risk value and the third risk value respectively; Compare and analyze the user's comprehensive risk assessment value with the preset reference risk interval. When the user's comprehensive risk assessment value is greater than the maximum value of the reference risk interval, the user is determined to be in a Class I risk state. When the user's comprehensive risk assessment value is within the reference risk interval, the user is determined to be in a Class II risk state. When the user's comprehensive risk assessment value is less than the minimum value of the reference risk interval, the user is determined to be in a Class III risk state. The risk type status is composed of Class I risk status, Class II risk status and Class III risk status.

9. The safe rental crowd evaluation system for a product rental platform according to claim 1 is characterized in that: Determine the target feature information interval. The specific process is as follows: Count the users who are judged to be in a certain risk state within a period of time to obtain a set of high-risk users; Count the users who are judged as Class II or Class III risk status within a period of time to obtain a set of low-risk users; The characteristic information of the corresponding users is extracted from the high-risk user set and the low-risk user set respectively. Multiple intervals are set for each characteristic information, and the characteristic information of each user is matched with the corresponding interval. Then, statistics are performed on each characteristic information interval, and the number of users in each interval is calculated. The characteristic information interval with the largest number of users is identified as the target characteristic information interval.

10. The safe rental crowd evaluation system for a product rental platform according to claim 1, characterized in that: It includes a verification and analysis module, which is used to verify and analyze the delineation results. The specific process is as follows: By obtaining the damage rate, overdue rate and revenue within a period of time after the demarcation results, the verification value is determined and compared with the set reference verification range for analysis; If the verification value is within the set reference verification range, a verification pass signal is generated; If the verification value is outside the set reference verification range, a verification failure signal is generated and the user's re-demarcation is triggered. During the re-demarcation process, the user's demarcation results will be adjusted according to the new analysis results, and verification analysis will be performed again until a verification pass signal is generated.

Images