Search

CN-122022963-A - Lease term management method and device

CN122022963ACN 122022963 ACN122022963 ACN 122022963ACN-122022963-A

Abstract

The application provides a lease period management method and device, relates to the technical field of lease management, and solves the technical problems that lease period management is stiff and cannot be dynamically adjusted according to equipment value dynamic evaluation and user behavior prediction in the prior art. The method comprises the steps of obtaining equipment body data, user portrait data and real-time behavior data, calculating a theoretical maximum safety period through a dynamic value attenuation model based on the equipment body data, predicting lease overage probability of a user through a long-short-period memory network based on the user portrait data and the real-time behavior data, and dynamically adjusting the current lease period, equipment function authority or triggering a buying instruction according to the lease overage probability and the equipment damage probability. The application is used for lease period management.

Inventors

  • Song Chenwu
  • LIU JUNJUN

Assignees

  • 合肥九九猪智能科技有限公司

Dates

Publication Date
20260512
Application Date
20260408

Claims (10)

  1. 1. A rental term management method, comprising: Acquiring equipment body data, user portrait data and real-time behavior data; calculating a theoretical maximum safety period through a dynamic value attenuation model based on the equipment body data; Predicting the lease expiration probability and the equipment damage probability of the user through a long-short-term memory network based on the user portrait data and the real-time behavior data; and dynamically adjusting the current lease period, the equipment function authority or the trigger buying instruction according to the grades of the lease exceeding probability and the equipment damage probability.
  2. 2. The method of claim 1, wherein the dynamic value decay model satisfies the following formula: Wherein, the For the current value of the device, For the initial value of the device, For the device reference devaluation coefficient, The correction coefficient is supplied and required for the market, The theoretical maximum safety period is as follows Is used for the time period of maximum duration of (a), Is a preset risk threshold.
  3. 3. The method of claim 1, wherein after dynamically adjusting a current rental period, device function authority, or trigger buy-off instruction according to the ranking of the rental period probability and the device damage probability, the method further comprises: constructing a credit score pool based on the historical performance of the user, and calculating a default Jin Jianmian coefficient or an overdue grace coefficient according to score data; and performing an elastic solution mechanism and an elastic grace mechanism based on the default Jin Jianmian coefficient and the overdue grace coefficient.
  4. 4. The method of claim 3, wherein the calculating the default Jin Jianmian coefficients or the overdue grace coefficients from the integral data satisfies the following formula: Wherein, the For the violations of the factor Jin Jianmian, For the overdue grace factor, And (5) judging the lease time exceeding probability.
  5. 5. The method according to claim 1, wherein predicting the rental probability and the equipment damage probability of the user through the long-term memory network based on the user portrait data and the real-time behavior data comprises: Determining a time sequence length k, and extracting the user portrait data and the real-time behavior data in the previous k days as a model input feature set; The input feature set is standardized and then input into a long-period memory network after training, wherein the long-period memory network after training takes historical user behavior data, performance data and equipment damage data of a leasing platform as a training set, and after the training set, a verification set and a test set are divided, the loss is calculated by adopting a cross entropy loss function, and the Adam optimizer is subjected to iterative optimization, and the long-period memory network is composed of two layers of long-period memory networks, a full-connection layer and a Sigmoid activation function, and is trained until the long-period memory network reaches the preset prediction accuracy; And outputting the rental expiration probability value and the equipment damage probability value in a preset prediction period of the user through an output layer of the long-short-period memory network.
  6. 6. The method of claim 1, wherein dynamically adjusting a current rental period, device function authority, or trigger buy-out instruction based on the ranking of the rental period probability and the device damage probability comprises: Presetting a first probability threshold and a second probability threshold, wherein the first probability threshold is smaller than the second probability threshold, and carrying out grading judgment on the lease overtime probability and the equipment damage probability based on the first probability threshold and the second probability threshold and executing corresponding operations: if the lease exceeding probability and the equipment damage probability are not greater than the first probability threshold, maintaining the current lease period and all function rights of the equipment; If the lease exceeding probability or the equipment damage probability is larger than the first probability threshold and not larger than the second probability threshold, the remaining lease period is adjusted to the end of the current charging period, the non-core function authority of the equipment is limited, and an air control early warning instruction is issued to the user terminal; if the lease exceeding probability is larger than a second probability threshold, shortening the remaining lease period to a preset short period number of days, limiting the high-value function permission of the equipment, and issuing a short period selection instruction to the user terminal; If the equipment damage probability is larger than a second probability threshold, the residual lease period is forcedly shortened to a preset buy-off consideration day, an equipment return or buy-off selection instruction is issued to the user terminal, if the feedback instruction of the user terminal is not received within the preset time, the equipment function is locked in a grading manner, and the forced buy-off instruction is issued to the user terminal.
  7. 7. The method according to claim 2, wherein the market supply and demand correction coefficient M (t) is calculated based on the transaction amount and the listing price fluctuation data of the same type of equipment in the second hand transaction platform in real-time docking, when the market supply and demand of the same type of equipment is not required, M (t) >1, when the market supply and demand are required, M (t) <1, and M (t) is updated daily; The physical loss coefficient S (t) is obtained by comprehensively calculating based on a boot self-checking report of the leased electronic product and hardware state data acquired by the Internet of things module, wherein the hardware state data comprises battery health, screen integrity, appearance color difference, hardware fault records and abnormal data of an acceleration sensor, the initial value of S (t) is 1, the higher the physical loss degree of the equipment is, the smaller the value of S (t) is, and the value range of S (t) is 0<S (t) is less than or equal to 1.
  8. 8. The method of claim 3, wherein the constructing a credit pool based on the user's historical performance comprises: Presetting an integral acquisition rule, an integral deduction rule and an integral accumulation rule based on user leasing performance data, and constructing a credit integral pool uniquely bound with a user main body; the point acquisition rule comprises the steps of counting the behavior of a user paying leases in time and quantity according to days, issuing basic points, issuing compliance points when detecting the illegal use behavior of the user without equipment, issuing performance rewards points when detecting the non-leasing risk early warning behavior of the user after the lease period is over, and issuing buying rewards points when detecting the behavior of the user buying off equipment in advance; The integral deduction rule comprises deducting corresponding integral according to overdue time when overdue behaviors of a user are detected, deducting preset integral at one time when jail/root of user equipment is exceeded and illegal operation behaviors of unauthorized dismounting are detected, and deducting corresponding integral according to the degree of damage in a grading manner when artificial damage behaviors of the equipment are detected; the point accumulation rule comprises the steps of setting effective duration for acquired points, accumulating the points in a cross-lease order, and storing point data in association with credit information of a user main body.
  9. 9. The method of claim 3 wherein the elastic reduction mechanism includes calculating a value of a factor α of a default Jin Jianmian after receiving an advance return application from a user, presetting a first threshold value of α, automatically exempting all of the default funds and executing an advance reduction instruction if α is greater than the first threshold value of α, thawing all of the deposit from the user, marking the device as a premium return asset, reducing the default funds by a ratio of the value of α if α is greater than a second threshold value of α and not greater than the first threshold value of α, executing an advance reduction instruction after the user pays the remaining default funds and thawing the remaining deposit if α is not greater than the second threshold value of α, executing an advance reduction instruction after withholding the default funds according to the original lease protocol if α is not greater than the second threshold value of α, wherein the second threshold value of α is less than the first threshold value of α; The elastic grace mechanism comprises the steps of calculating the value of a overdue grace coefficient beta after detecting overdue behaviors of a user, presetting a beta first threshold, automatically giving a first preset grace period if beta is larger than the beta first threshold, giving no overdue penalty information and no limit on equipment function authority in the grace period, only issuing a refund reminding instruction to the user terminal, giving a second preset grace period if beta is larger than a preset beta second threshold and is not larger than the beta first threshold, giving a refund reminding instruction for issuing a tape default interest to the user terminal, and immediately starting normal penalty information calculation and issuing a overdue wind control early warning instruction to the user terminal if beta is not larger than the beta second threshold, wherein the beta second threshold is smaller than the beta first threshold, and the second preset grace period is shorter than the first preset grace period.
  10. 10. A lease term management apparatus is characterized in that the apparatus includes a communication unit and a processing unit; The communication unit is used for acquiring equipment body data, user portrait data and real-time behavior data; The processing unit is used for calculating a theoretical maximum safety period through a dynamic value attenuation model based on the equipment body data, predicting the lease expiration probability and the equipment damage probability of a user through a long-short-period memory network based on the user portrait data and the real-time behavior data, and dynamically adjusting the current lease period, the equipment function authority or triggering a purchase order according to the lease expiration probability and the equipment damage probability.

Description

Lease term management method and device Technical Field The application relates to the technical field of lease management, in particular to a lease period management method and a lease period management device. Background Electronic product renting has become an important business model in the consumer electronics field. Currently, in the field of electronic product renting, a mode of fixed renting period is generally adopted, and after the renting period expires, a user needs to manually return, renew renting or buy off. However, the electronic product has the characteristics of high depreciation rate and high liquidity, the real-time residual value of the equipment cannot be reflected by the fixed term, the temporary equipment value in the aspect of renting is lower than the credit risk of the residual rent, meanwhile, the fixed lease cannot be matched with the personalized service period of the user, the loss and the order conversion rate of the user are lower, the dynamic adjustment mechanism based on the real-time use behavior of the user is lacking in the prior art, the renting authority or deposit threshold cannot be dynamically adjusted according to the real-time behavior of the user, the default disposal cost is high, the hierarchical disposal mechanism is lacking, and the asset operation efficiency and dispute resolution efficiency are low. Therefore, the prior art has the technical problems that the management of the lease term is stiff and the lease term cannot be dynamically adjusted according to the equipment value dynamic evaluation and the user behavior prediction. Disclosure of Invention The application provides a lease period management method and a lease period management device, which solve the technical problems that lease period management is stiff and dynamic adjustment cannot be carried out according to equipment value dynamic evaluation and user behavior prediction in the prior art. In order to achieve the above purpose, the application adopts the following technical scheme: The first aspect provides a lease term management method, which comprises the steps of obtaining equipment body data, user portrait data and real-time behavior data, calculating a theoretical maximum safety period through a dynamic value attenuation model based on the equipment body data, predicting lease overage probability and equipment damage probability of a user through a long-short-term memory network based on the user portrait data and the real-time behavior data, and dynamically adjusting current lease term, equipment function permission or triggering a buying instruction according to the lease overage probability and the equipment damage probability. With reference to the first aspect, in one possible implementation manner, the dynamic value attenuation model satisfies the following formula: Wherein, the For the current value of the device,For the initial value of the device,For the device reference devaluation coefficient,The correction coefficient is supplied and required for the market,The theoretical maximum safety period is as followsIs used for the time period of maximum duration of (a),Is a preset risk threshold. With reference to the first aspect, in one possible implementation manner, after dynamically adjusting the current rental period, the device function authority, or triggering the buy-out instruction according to the classification of the rental period exceeding probability and the device damage probability, the method further includes constructing a credit pool based on the historical performance of the user, calculating a default Jin Jianmian coefficient or a overdue grace coefficient according to the integral data, and performing an elastic solution mechanism and an elastic grace mechanism based on the default grace coefficient and the overdue grace coefficient. With reference to the first aspect, in one possible implementation manner, the calculating of the default Jin Jianmian coefficient or the overdue grace coefficient according to the integral data satisfies the following formula: Wherein, the In order to violate the factor Jin Jianmian,In order to be a overdue grace factor,And (5) the lease time exceeding probability is used for lease. With reference to the first aspect, in one possible implementation manner, the prediction of the rental period probability and the equipment damage probability of the user through the long-short-period memory network based on the user portrait data and the real-time behavior data includes determining a time sequence length k, extracting the user portrait data and the real-time behavior data of the previous k days as model input feature sets, performing standardization processing on the input feature sets, inputting the standardized input feature sets into the trained long-short-period memory network, wherein the trained long-short-period memory network uses the historical user behavior data, the performance data and the equipment damage data of