Search

CN-122022961-A - Big data-based analysis and management method for car rental credit information

CN122022961ACN 122022961 ACN122022961 ACN 122022961ACN-122022961-A

Abstract

The invention relates to the technical field of car leasing and discloses a car leasing credit information analysis and management method based on big data, which comprises the following steps of firstly, collecting operation data of a network vehicle driver, wherein the operation data comprise daily order quantity, nutrient composition, platform punishment, customer service complaints and traffic violation records; the method comprises the steps of collecting asset registration information of a leased vehicle, wherein the asset registration information comprises vehicle unique identification information, property persons, leasing contract records, mortgage or mortgage records and insurance application information, and constructing a network-bound vehicle driver, a leased vehicle and a leasing contract based on collected operation data and asset registration information. By adopting the technical scheme of double-time-axis credit index extraction and dynamic credit curve calculation, the technical effects of tracking the operation of a driver and the dynamic change of an asset in real time are achieved, the continuous and accurate prediction of the performance capability of a network about vehicle driver is realized, and the defects that the prior art depends on static information and cannot effectively predict the real repayment capability are overcome.

Inventors

  • FU YUAN

Assignees

  • 北京首成佳茂科技有限公司

Dates

Publication Date
20260512
Application Date
20260203

Claims (10)

  1. 1. The analysis and management method for the car rental credit information based on the big data is characterized by comprising the following steps of: Collecting operation data of a network taxi driver, including daily order quantity, nutrient composition, platform punishment, customer service complaints and traffic violation records, and collecting asset registration information of a lease vehicle, including vehicle unique identification information, property persons, lease contract records, mortgage or mortgage records and insurance application information; step two, constructing a multi-subject relationship map taking a net car driver, a leasing car, a leasing contract and a fund provider as nodes based on the collected operation data and asset registration information; Extracting double-time-axis credit indexes reflecting the performance capability of a network about vehicle driver based on the constructed multi-subject relation map, wherein the double-time-axis credit indexes comprise operation time-axis indexes reflecting the operation stability of the driver and adverse events, and asset time-axis indexes reflecting the occupied integrity, the performance consistency and abnormal registration of vehicle assets; Calculating a dynamic credit curve of a network taxi driver changing in a lease period in real time by using a preset weight model according to the extracted double-time-axis credit index, wherein the dynamic credit curve is used for forming continuous credit records and risk images; Step five, executing a risk linkage control strategy according to the change trend of the dynamic credit curve and the reached preset threshold, wherein the risk linkage control strategy comprises pre-warning triggered when the dynamic credit curve continuously descends, limiting measures or strategy adjustment executed when the dynamic credit curve reaches the low threshold, and rewarding level preference given when the dynamic credit curve is in a high position for a long time; and step six, generating a unified credit analysis visual view according to the dynamic credit curve and the multi-subject relation graph, and providing credit grade information after authorization and desensitization for an external association platform.
  2. 2. The method for analyzing and managing car rental credit information based on big data according to claim 1, wherein the second step further comprises: Substep Based on the collected operational data and asset registration information, a set of principal nodes of a multi-principal relational graph is defined And entity attribute set The subject node set Comprising network bus driver nodes Vehicle node Contract node Fund side node The entity attribute set Storing attribute values of all nodes; Substep Defining the set of principal nodes based on the collected operational data and asset registration information Relationship edge set Each relationship edge according to relationship type and asset importance Calculating initial trust weights The initial trust weight The calculation formula of (2) is as follows: , Wherein, the Is a relationship edge Is used to determine the initial trust weight of (1), The function is indicated for the type of relationship, For a cost function that relates the assets involved, And Assigning a factor to a preset weight, and The method is used for balancing the influence of the relationship type and the asset value on the trust weight; Substep Judging the relation edge set Whether or not there are multiple relationship edges If the situation of the same vehicle node is pointed at, judging that the vehicle node is in a multiple occupancy risk state, and calculating a multiple occupancy risk index of the vehicle node according to the following formula: , Wherein, the For vehicle nodes Is a multiple occupancy risk index of (2), To point to a vehicle node Is a subset of all of the relationship edges of (a), Is a relationship edge Is used to determine the initial trust weight of (1), Indicating factors for relationship conflicts; Aggregating principal nodes Relationship edge set And multiple occupancy risk index Commonly stored to form a multi-subject relationship graph 。
  3. 3. The method for analyzing and managing car rental credit information based on big data according to claim 1, wherein the third step further comprises: Substep Calculating an operation time axis index based on the collected operation data Comprises a composition factor of average daily income Stability of operation Adverse event index ; The average daily income Obtained by average calculation of daily order campaigns in a period, the operation stability Obtained by evaluating the fluctuation degree of daily income in a period or the consistency of operation time length, wherein the adverse event index The severity of platform punishment, major traffic violations and customer service complaints in the period is obtained in an accumulated manner; using linear normalizing functions Normalizing the composition factors, wherein the operation time axis index The calculation formula of (2) is as follows: , Wherein, the Is time of The operation time axis index of the time is integrated into a score, Is time of The daily average income of the time is calculated, Is time of The operational stability at the time of the operation, Is time of An adverse event index at the time of the test, 、 And Is a preset operation index weight used for reflecting the contribution degree of each composition factor to the operation credit, and ; Substep Based on asset registration information and multi-principal relationship atlas Calculating asset timeline index Is used as a component factor of the (c) in the composition, including asset occupancy integrity Consistency of performance Abnormality registration index ; The asset occupancy integrity By multiple occupancy risk indices Reverse mapping is carried out to obtain the degree of locking of the vehicle assets by multiple parties, and the consistency of the performance is reflected The abnormal registration index is obtained by analyzing the deviation times and the amounts of the rents paid by the drivers in time and the contracted amounts in the period The method comprises the steps of accumulating and obtaining the number and the type of abnormal records of the vehicle assets on a public query system or an industry sharing platform; Using the linear normalization function Normalizing the composition factors to obtain the asset time axis index The calculation formula of (2) is as follows: , Wherein, the Is time of The asset timeline index at that time is a composite score, Is time of The assets at the time occupy the integrity of the system, Is time of The consistency of the performance of the vehicle is achieved, Is time of An abnormal registration index at the time of the time, 、 And Is a preset asset index weight used for reflecting the contribution degree of each composition factor to the asset credit, and ; Substep Will operate the time axis index And asset timeline index As the output result of the double time axis credit index, and at the same time, the operation time axis index And asset timeline index Performing logic verification if the operation time axis index is Continuous and continuous Operation warning line with period lower than preset period And asset timeline index Asset warning line below preset Triggering a primary credit corrupted signal ; The output result operates the time axis index And asset timeline index Will be used for dynamic credit curve calculation.
  4. 4. The method for analyzing and managing car rental credit information based on big data according to claim 1, wherein the fourth step further comprises: Substep Operation time axis index based on output And asset timeline index Calculate the time Dynamic credit initial score at time The dynamic credit initial score The calculation formula of (2) is as follows: , Wherein, the Is time of An initial score of the dynamic credit at the time, For a preset operational timeline weight, The method comprises the steps of (1) weighing a preset asset time axis; the dynamic credit initial score Will be used for credit smoothing correction; Substep Initial score based on dynamic credit Performing exponential smoothing correction to obtain time Dynamic credit curve score for time The dynamic credit curve score The calculation formula of (2) is as follows: , Wherein, the Is time of The dynamic credit curve score at the time of the time, For a dynamic credit initial score, For the last time The dynamic credit curve score at the time of the time, Is a preset smoothing factor; at the same time by continuously Periodic dynamic credit curve score Performing linear fitting to calculate credit change slope For reflecting the instant trend of credit; Substep Dynamic credit curve based score And credit change slope Executing multidimensional credit risk level judgment and taking time Credit status classification into risk classes The risk level The determination conditions of (2) are: , Wherein, the Is time of A level of credit risk at the time of the transaction, 、 And For a preset threshold value of the credit rating, Is a preset rapid deterioration slope threshold; The dynamic credit curve score Slope of credit change And risk level And as a result of the output of the dynamic credit curve, the method is used for risk linkage control.
  5. 5. The method for analyzing and managing car rental credit information based on big data according to claim 1, wherein the fifth step further comprises: Substep Receiving an output dynamic credit curve score And credit change slope Combining primary credit corrupted signals Executing pre-early warning judgment, wherein the pre-early warning judgment condition The method comprises the following steps: , Wherein, the In order to pre-warn the judgment condition in advance, For the slope of the credit change, Is a preset early warning slope threshold value, Is a preset early warning score threshold value, Is a primary credit corrupted signal; If the condition is early-warning in advance In order to establish, the pre-warning notice to leasing company and network taxi driver is automatically triggered, and the notice is recorded to the constructed multi-subject relation map In (a) and (b); Substep Continuously monitoring the dynamic credit curve score after triggering the pre-alarm notification If dynamic credit curve score Continuous after early warning notification The early warning score threshold value is still lower in the period Or risk level Is at high risk Automatically executing risk limiting measures Policy adjustment The risk limiting measures Execution determination condition of (2) The method comprises the following steps: , Wherein, the For the execution determination condition of the risk limiting measure, Is a preset number of duration cycles; the risk limiting measures of the execution Policy adjustment Will be the result of this linkage control; Substep Dynamic credit curve based score And risk level Executing the judgment of the bonus level offers, wherein the judgment conditions of the execution of the bonus level offers The method comprises the following steps: , Wherein, the A decision condition for execution of the bonus level benefit, For the average value of dynamic credit curve scores of net-bound car drivers in the investigation period, For a preset average score threshold value of the prize, The risk grade of the network vehicle driver is high in quality in the investigation period Is used for the time-to-time ratio of (c), A preset high-quality time duty ratio threshold value; if the execution judgment condition of the bonus level preference And (3) establishing the automatic awarding of the online taxi driver with the bonus-level offer, wherein the bonus-level offer comprises the steps of reducing the deposit proportion of the renewing lease or new lease contract and prolonging the repayment period, and the bonus-level offer is directly reflected as credit asset deposit.
  6. 6. The big data-based analysis and management method for car rental credit information according to claim 2, wherein the multi-subject relationship map is characterized in that Adopting a graph database to store, wherein the storage structure of the graph database stores the driver nodes Vehicle node Contract node Fund side node Attribute set Stored as nodes and collecting the relationship edges Stored as directed edges that simultaneously store initial trust weights 。
  7. 7. The big data based analysis and management method for car rental credit information according to claim 5, wherein the execution frequency of the risk coordinated control strategy is dynamically adjusted, and the execution frequency is The adjustment determination conditions of (2) are: , Wherein, the As the execution frequency of the risk linked control strategy, As a risk level of the risk level, In order to pre-warn the judgment condition in advance, 、 And High, medium and low frequencies, and meets > > For achieving high frequency intervention for high risk states.
  8. 8. A big data based car rental credit analysis and management method as in claim 3, wherein the asset occupancy integrity By multiple occupancy risk indices The functional relation for reverse mapping is: , Wherein, the For the purpose of asset occupancy integrity, Is a multiple occupancy risk index; Is a monotonically increasing nonlinear mapping function for indexing multiple occupancy risks The map is the degree of loss of integrity of the asset occupancy, and the higher the risk index, the lower the integrity score.
  9. 9. The method for analyzing and managing car rental credit information based on big data as claimed in claim 3, wherein the adverse event index is Is calculated by adopting a weighted summation mode: , Wherein, the In order to be an adverse event index, For the total number of adverse events occurring in a cycle, Is the first A preset penalty weight for the adverse event, Is the first The quantitative severity factor of the adverse event is as follows: when the adverse event is a platform penalty, A penalty rank score applied to the platform; when the adverse event is a traffic violation, Cumulative score or fine amount recorded for the traffic management; When the adverse event is a customer service complaint, The score is rated for the severity of the complaint after verification.
  10. 10. The method for analyzing and managing big data-based car rental credit information according to claim 1, wherein the collection of the operation data of the network about car driver and the asset registration information of the rental car is done in streaming real-time or near real-time, and the real-time collection frequency of the operation data is not lower than that of the operation data The quasi-real-time acquisition frequency of the asset registration information is not lower than Times/day, ensure dynamic credit curve The timeliness of the basic data is calculated; Wherein, the For ensuring the operation time axis index for the preset minimum acquisition frequency of the operation data Real-time performance of calculation; Registering the lowest acquisition frequency of information for a preset asset for ensuring the time axis index of the asset And (5) calculating quasi-real time performance.

Description

Big data-based analysis and management method for car rental credit information Technical Field The invention relates to the technical field of car leasing, in particular to a car leasing credit information analysis and management method based on big data. Background The rapid development of mobile travel and shared economy, and car renting, especially car renting and financing renting businesses for network-bound car drivers, has become an important component in car finance and travel service ecology. Unlike traditional personal credit or consumer finance, internet-oriented car rental business for car drivers has unique complexity and high risk. In the prior art, in the assessment of car rental credit, static information such as income evidence, credit report, bank running water and the like before signing is relied on, and continuous tracking of repayment capability indexes such as dynamic income, oil cost road expense, family liability and the like in the actual operation process of a network taxi driver is lacked, so that the real repayment capability of the driver under market fluctuation or policy adjustment cannot be effectively predicted, and overdue risks of the surface qualified driver in the follow-up operation are extremely easy to occur. In the existing car rental credit management flow, risk signals such as the decrease of the receipt quantity, the punishment of a platform, the score of traffic violations, the increase of complaints and the like generated by a driver in the operation process are stored in different systems in a scattered mode, and a rental company is not easy to integrate discrete information to form continuous credit records and risk portraits, so that risk early warning is delayed and an effective process intervention mechanism is lacked. The existing car leasing business relates to multiple participators such as car manufacturers, financial leasing companies, network taxi-booking platforms and the like, credit information among the multiple participators lacks transparent and uniform sharing and linkage mechanisms, a uniform credit view of a vehicle and a driver cannot be formed, and risks of information cracking, unclear responsibility and even 'multi-taxi' in leasing transaction are easily caused. In the existing car rental system, a good network car-closing driver who obeys a rental contract for a long time, keeps good driving habit and pays full amount on time can not be effectively proved and transferred across a main body when a cooperative platform or a rental company is replaced by accumulated good credit records, so that the good network car-closing driver is forced to bear high deposit and risk clauses again, the operation cost is increased, and meanwhile, the capability of the rental company for attracting and screening high-quality clients is limited. In order to solve the problems, the invention provides a car rental credit information analysis and management method based on big data. According to the method, a multi-main-body relation graph is constructed by collecting multi-source data, and a double-time-axis dynamic credit curve is calculated in real time, so that continuous prediction of the performance capability of a net-bound vehicle driver and dynamic linkage control of risks are realized. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a car rental credit information analysis and management method based on big data, which aims to solve the problems in the prior art. In order to achieve the aim, the invention is realized by the following technical scheme that the automobile leasing credit information analysis and management method based on big data comprises the following steps: Collecting operation data of a network taxi driver, including daily order quantity, nutrient composition, platform punishment, customer service complaints and traffic violation records, and collecting asset registration information of a lease vehicle, including vehicle unique identification information, property persons, lease contract records, mortgage or mortgage records and insurance application information; step two, constructing a multi-subject relationship map taking a net car driver, a leasing car, a leasing contract and a fund provider as nodes based on the collected operation data and asset registration information; Extracting double-time-axis credit indexes reflecting the performance capability of a network about vehicle driver based on the constructed multi-subject relation map, wherein the double-time-axis credit indexes comprise operation time-axis indexes reflecting the operation stability of the driver and adverse events, and asset time-axis indexes reflecting the occupied integrity, the performance consistency and abnormal registration of vehicle assets; Calculating a dynamic credit curve of a network taxi driver changing in a lease period in real time by using a preset weight model according to the extracted double-time-axi