Driver empty driving compensation system based on online car-hailing order cancellation scene

Abstract

The invention discloses a driver empty driving compensation system based on an online car-hailing order cancellation scene, which relates to the field of intelligent travel services and comprises a cancellation analysis module, a responsibility judgment analysis module, an empty driving compensation module and an appeal management module. The order information is automatically collected, the NLP technology is utilized to analyze and cancel reasons, the preliminary judgment rule base is configured, the responsibilities of passengers and drivers are defined, and the responsibility judgment process is optimized through the Gaussian mixture model. Setting a lowest unloaded distance standard and a compensation period, limiting the number of application times of a driver, and effectively managing compensation application; setting an empty driving compensation standard, defining compensation conditions and compensation times, and calculating total compensation cost according to the basic compensation cost, the kilometer compensation cost and the to-be-supplemented and repaid cost; and for the non-compensation condition, the appeal is submitted through the mobile application and audited and fed back within 48 hours, and the appeal record is stored, so that the transparency and the user trust are improved, the operation experience of the driver can be optimized, the dispute is reduced, and the compensation auditing efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent travel services, and in particular to a driver idle driving compensation system based on a scenario where an online car-hailing order is cancelled. Background Art

[0002] With the increasing number of service providers integrated by the Bailongma SASS system, the number of drivers and orders has increased, and the problem of drivers suffering losses after orders are canceled has become increasingly serious. When picking up passengers, online car-hailing drivers often encounter passengers who suddenly cancel orders for various reasons. Drivers will have empty driving costs for picking up passengers, and the need for empty driving compensation is becoming more and more urgent.

[0003] From the perspective of matching drivers and passengers, there are uncertainties in long-distance orders. On the one hand, drivers dare not pick up passengers, fearing that the passengers will cancel their rides and waste fuel due to empty trips. On the other hand, passengers are unwilling to wait and worry that the driver will not be able to pick them up on time. From the perspective of driver retention, there is the issue of compensation for drivers' empty trips. From the perspective of service provider costs, some service providers currently subsidize drivers offline to ensure driver retention, but the verification and distribution costs are high, and the efficiency is low.

[0004] The existing technology has the following deficiencies: In the online car-hailing industry, the traditional idle-trip compensation system has a relatively vague responsibility determination, and often cannot accurately divide the responsibilities of passengers and drivers, resulting in different compensation standards; the unclear compensation rules make drivers confused when applying for compensation, and they cannot clearly understand whether they meet the compensation conditions; in the traditional system, compensation applications often require a long time of manual review, causing drivers to wait a lot. The opacity of compensation information makes drivers lack trust in the compensation calculation process. Insufficient data processing capabilities cannot efficiently process large amounts of cancellation data.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0006] The purpose of the present invention is to provide a driver idle driving compensation system based on the online car-hailing order cancellation scenario to solve the problems in the above-mentioned background technology.

[0007] In order to achieve the above-mentioned purpose, the present invention provides the following technical solutions: a driver idle driving compensation system based on the online car-hailing order cancellation scenario, comprising a cancellation analysis module, a responsibility analysis module, an idle driving compensation module and a complaint management module;

[0008] Cancellation analysis module: collects basic order information and automatically inputs it into the penalty model, divides the order cancellation party into passenger cancellation party and driver cancellation party and marks the cancellation ID, and predefines the cancellation reason options of the cancellation ID according to the cancellation action;

[0009] Responsibility analysis module: configures a preliminary judgment rule base, builds a Gaussian mixture model, and combines driver feedback and customer service follow-up results to make a probabilistic judgment on the responsibility for order cancellation;

[0010] Idle driving compensation module: set idle driving compensation standards, define compensation conditions and compensation times, calculate total compensation fees based on basic compensation fees, kilometer compensation fees and waiting compensation fees, and display, monitor and record abnormal applications;

[0011] Complaint management module: Provide drivers with an online complaint channel through mobile applications, record the reasons and evidence for complaints, set review time limits and feedback results, and transparently display compensation applications and complaint status.

[0012] Preferably, after the order is canceled, the system collects basic information of the order and automatically inputs it into the judgment model. The system log is used to retrieve the order canceling party information specified in the order, and the order canceling party is divided into passenger canceling party and driver canceling party and the cancellation ID is marked. The cancellation reason options of the cancellation ID are predefined according to the cancellation action. If the cancellation reason is detected as free text input, NLP technology is used to analyze the input text of the cancellation ID and identify keywords to classify the cancellation reasons.

[0013] Preferably, a preliminary judgment rule base is configured. When the cancellation behavior is caused by the passenger and the cancellation reason is reasonable, the driver is judged to be not responsible. When the cancellation behavior is caused by the driver and the cancellation reason is unreasonable, the driver is judged to be responsible. When it is impossible to judge, the system selects the judgment state, and constructs a Gaussian mixture model to identify the judgment of different dimensions of the judgment state. The number of Gaussian components in the dimension of driver feedback cancellation reason and the dimension of customer service return visit passenger result are selected through cross-validation, and each Gaussian component number and covariance matrix are randomly initialized. The probability of responsibility allocation corresponding to the number of Gaussian components in each dimension of driver feedback cancellation reason and customer service return visit passenger result is calculated. The specific formula is:

[0014]

[0015] Among them, γ [(i,j)(k,l)] represents the probability of the responsibility distribution corresponding to Gaussian components k and j, respectively, for the driver's feedback cancellation reason dimension i and the customer service return visit passenger result dimension j, N(V i |μ k ,∑k) represents the probability density of the driver’s feedback cancellation reason dimension i under Gaussian distribution k, where V i Indicates the driver's feedback cancellation reason dimension i, μ k represents the mean parameter of the driver feedback cancellation reason dimension i under Gaussian distribution k, ∑k represents the covariance parameter of the driver feedback cancellation reason dimension i under Gaussian distribution k, N(I j |μ l,∑l) represents the probability density of the customer service return visit result dimension j corresponding to Gaussian distribution j, where I j represents the dimension j of the customer service return visit result, μ l represents the mean parameter of the customer service return visit result dimension j under Gaussian distribution I, ∑l represents the covariance parameter of the customer service return visit result dimension j under Gaussian distribution I, It represents the total probability density of all Gaussian components m, and updates the covariance parameters of each Gaussian component. The log-likelihood function is used to iteratively converge the identification and judgment of different dimensions of the state.

[0016] Preferably, the minimum idle driving distance standard is set to 2 kilometers, and the actual driving distance of the driver from the time the order is accepted to the time the driver cancels the order is recorded and compared with the distances between the starting point and the alighting point set by the passenger. The actual route of the driver is compared with the preset navigation route in the system. If the actual route of the driver deviates from the preset navigation route by 40%, it is judged that the compensation conditions are not met. The compensation period is set to one month, and the number of idle driving compensation that the driver can apply for within the compensation period is limited to 5 times. The system tracks the number of idle driving compensation that the driver has applied for within the compensation period. If the upper limit of the number of idle driving compensation is reached, the driver is not allowed to submit a new compensation application. The time when the driver feeds back "waiting for passengers" to the system is recorded. If there is a significant gap between the time fed back by the driver and the estimated arrival time of the navigation, it is judged that the system feedback is unreasonable and no compensation is given. The GPS system is connected to determine whether the driver is close to the boarding point. Within a distance range with a radius of 3m, the system feedback is judged to be reasonable and compensation is given. A compensation rule feedback mechanism is established, and the driver's compensation application history is monitored, and abnormal compensation applications are marked.

[0017] Preferably, a fixed basic compensation fee is defined to compensate all empty orders that meet the conditions, a kilometer compensation fee standard is defined, the time the driver waits for passengers to board the vehicle after arriving at the pick-up point is recorded and a waiting compensation fee is defined, the total compensation fee is calculated based on the basic compensation fee, kilometer compensation fee and waiting compensation fee, the total compensation fee calculation result is displayed to the driver through a visualization platform, and the details of each compensation calculation are recorded in the system compensation log.

[0018] Preferably, the driver receives a no-compensation notification determined by the system, submits reasons for appeal against the no-compensation result, sets up a driver appeal account through the online appeal channel of the mobile application, adds an idle-mile compensation option to fill in the driver's appeal reasons, and notifies the driver to attach driving record evidence including driving route and driving time, passenger conversation record evidence, and emergency evidence including traffic and weather conditions. The system accepts and records all information in the driver's appeal, sets a 48-hour review period to complete the review and feedback of the appeal, and if the appeal is approved, the compensation amount is recalculated and notified to the driver; if the appeal is not approved, the system provides clear reasons and relevant terms, saves all information of each driver's appeal to the system appeal log, and sets an idle-mile compensation dashboard to transparently display the driver's compensation application and appeal status, the status of each appeal, the reasons for the appeal and the summary of evidence.

[0019] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0020] 1. By integrating the basic information of the order, the driver's feedback, the customer service follow-up results and the Gaussian mixture model, an in-depth analysis is performed using advanced statistical learning methods. Specifically, the Gaussian mixture model makes a probabilistic judgment on the responsibility for order cancellation by analyzing the cancellation reasons reported by the driver and the information from the customer service follow-up. Combining multiple data dimensions together to generate a multi-dimensional probability distribution on the attribution of responsibility can comprehensively consider various factors, eliminate the judgment bias caused by single-source information, and achieve more accurate judgment. Continuously optimize in actual operation, improve the accuracy of the model through continuous updates, and be able to monitor the cancellation behavior performance in different situations in real time, help the operation team identify potential problem areas, and then take targeted measures. By improving the accuracy of the cancellation reason analysis, the trust relationship between drivers and passengers is enhanced, and the overall user experience is optimized. The transparency of the analysis process also makes passengers and drivers more confident in the system's judgment, reduces disputes caused by misunderstandings, creates travel convenience for users, and improves the competitiveness of the platform.

[0021] 2. Reduce reliance on manual intervention by automating the more cumbersome steps of information collection, process management, and compensation application. Whenever a driver applies for compensation or makes a complaint, the relevant information is automatically recorded and a background check is performed to quickly review the application. By setting a 48-hour review period, it is ensured that all applications can receive timely feedback, and efficient response provides users with a good service experience. The compensation calculation results and complaint status are clearly displayed on the visual platform, allowing drivers to understand the compensation progress and complaint results, and improving transparency. Potential fraud or unreasonable compensation applications can be identified through intelligent marking of abnormal applications, effectively maintaining the fairness and impartiality of the platform. The need for manual intervention is reduced through automated processes, which also significantly reduces operating costs, improves the response speed of the entire order processing and feedback mechanism, and enhances user satisfaction and the market competitiveness of the platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0023] Figure 1 This is a module schematic diagram of a driver idle driving compensation system based on an online car-hailing order cancellation scenario according to the present invention. DETAILED DESCRIPTION

[0024] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0025] The present invention provides Figure 1 A driver idle driving compensation system based on the online car-hailing order cancellation scenario is shown, including a cancellation analysis module, a responsibility analysis module, an idle driving compensation module and a complaint management module;

[0026] Cancellation analysis module: collects basic order information and automatically inputs it into the penalty model, divides the order cancellation party into passenger cancellation party and driver cancellation party and marks the cancellation ID, and predefines the cancellation reason options of the cancellation ID according to the cancellation action;

[0027] After the order is canceled, the system collects basic information about the order and automatically inputs it into the judgment model, including order number, driver information, passenger information, cancellation time, and cancellation reason. The system log is used to retrieve the order canceling party information specified in the order, and the order canceling party is divided into passenger canceling party and driver canceling party, and the cancellation ID is marked. The cancellation reason options for the cancellation ID are predefined based on the cancellation action, including "temporary matters", "driver late", and "refusal to carry". If the cancellation reason is detected as free text input, NLP technology is used to analyze the input text of the cancellation ID and identify keywords to classify the cancellation reasons.

[0028] Responsibility analysis module: configures a preliminary judgment rule base, builds a Gaussian mixture model, and combines driver feedback and customer service follow-up results to make a probabilistic judgment on the responsibility for order cancellation;

[0029] Configure the preliminary judgment rule base. When the cancellation behavior is caused by the passenger and the cancellation reason is reasonable, the driver is judged to be not responsible. When the cancellation behavior is caused by the driver and the cancellation reason is unreasonable, the driver is judged to be responsible. When it is impossible to judge, the system selects the judgment state, and constructs a Gaussian mixture model to identify the judgment of different dimensions in the judgment state. The number of Gaussian components in the dimension of driver feedback cancellation reasons and the dimension of customer service return visit to passengers are selected through cross-validation. Each Gaussian component number and covariance matrix are randomly initialized, and the responsibility allocation probability of each Gaussian component number corresponding to each driver feedback cancellation reason dimension and customer service return visit to passengers result dimension is calculated. The specific formula is:

[0030]

[0031] Among them, γ [(i,j)(k,l)] represents the probability of the responsibility distribution corresponding to Gaussian components k and j, respectively, for the driver's feedback cancellation reason dimension i and the customer service return visit passenger result dimension j, N(V i |μ k ,∑k) represents the probability density of the driver’s feedback cancellation reason dimension i under Gaussian distribution k, where V i Indicates the driver's feedback cancellation reason dimension i, μ k represents the mean parameter of the driver feedback cancellation reason dimension i under Gaussian distribution k, ∑k represents the covariance parameter of the driver feedback cancellation reason dimension i under Gaussian distribution k, N(I j |μ l ,∑l) represents the probability density of the customer service return visit result dimension j corresponding to Gaussian distribution j, where I j represents the dimension j of the customer service return visit result, μ l represents the mean parameter of the customer service return visit result dimension j under Gaussian distribution I, ∑l represents the covariance parameter of the customer service return visit result dimension j under Gaussian distribution I, It represents the total probability density of all Gaussian components m, and updates the covariance parameters of each Gaussian component. The log-likelihood function is used to iteratively converge the identification and judgment of different dimensions of the state.

[0032] Idle driving compensation module: set idle driving compensation standards, define compensation conditions and compensation times, calculate total compensation fees based on basic compensation fees, kilometer compensation fees and waiting compensation fees, and display, monitor and record abnormal applications;

[0033] The minimum idle driving distance standard is set to 2 kilometers. The actual driving distance of the driver from the time the order is accepted to the time the driver cancels the order is recorded and compared with the distance between the starting point and the drop-off point set by the passenger. The actual route of the driver is compared with the preset navigation route in the system. The actual route of the driver deviates from the preset navigation route by 40%, which is judged to be ineligible for compensation. The compensation period is set to one month, and the number of idle driving compensation that the driver can apply for within the compensation period is limited to 5 times. The system tracks the number of idle driving compensation that the driver has applied for within the compensation period. If the upper limit of the number of idle driving compensation is reached, the driver is not allowed to submit a new compensation application. The time when the driver reports "waiting for passengers" is recorded and fed back to the system. There is a significant gap between the time fed back by the driver and the estimated arrival time of the navigation. The system feedback is judged to be unreasonable and no compensation is given. The GPS system is connected to determine whether the driver is close to the boarding point. Within a distance radius of 3m, the system feedback is judged to be reasonable and compensation is given. A compensation rule feedback mechanism is established, and the driver's compensation application history is monitored, and abnormal compensation applications are marked.

[0034] Define a fixed basic compensation fee to compensate all empty orders that meet the conditions, set the basic compensation fee to 10 yuan, define the kilometer compensation fee standard, set the kilometer compensation fee to 2 yuan per kilometer, record the time the driver waits for passengers to get on the bus after arriving at the pick-up point and define the waiting compensation fee, set the waiting compensation fee to 1 yuan per minute, calculate the total compensation fee based on the basic compensation fee, kilometer compensation fee and waiting compensation fee, display the total compensation fee calculation results to the driver through the visualization platform, and record the details of each compensation calculation in the system compensation log.

[0035] Complaint management module: Provide drivers with an online complaint channel through mobile applications, record the reasons and evidence for complaints, set review time limits and feedback results, and transparently display compensation applications and complaint status;

[0036] The driver receives a notification of no compensation determined by the system and submits reasons for appeal against the no compensation result. He or she sets up a driver appeal account through the online appeal channel of the mobile application, adds an option for idle mileage compensation, fills in the driver's reasons for appeal, and notifies the driver to attach driving record evidence including driving route and driving time, passenger conversation record evidence, and emergency evidence including traffic and weather conditions. The system accepts and records all information in the driver's appeal, sets a 48-hour review period to complete the review and feedback of the appeal, and if the appeal is approved, the compensation amount will be recalculated and notified to the driver. If the appeal is not approved, the system will provide clear reasons and relevant terms, save all information of each driver's appeal in the system appeal log, set up an idle mileage compensation dashboard, and transparently display the driver's compensation application and appeal, the status of each appeal, the reasons for the appeal and the summary of evidence.

[0037] The cancellation analysis module greatly improves the efficiency of analyzing order cancellation behavior by automatically collecting order information and cancellation data. The system automatically inputs order cancellation information into the penalty model, clearly distinguishes the cancellation responsibilities of passengers and drivers, and generates a unique cancellation ID for each cancellation event. This process not only reduces errors caused by manual operations, but also speeds up information processing. At the same time, according to the predefined cancellation reason options, the reasons for passenger or driver cancellation are effectively classified, ensuring the accuracy of subsequent responsibility determination. For the cancellation reasons entered in free text, the system uses natural language processing technology to automatically identify and classify keywords, thereby achieving efficient and accurate cancellation reason analysis. Through the information retrieved from the system log, the data processing of cancellation behavior makes the data transparent and traceable, and enhances the platform's trust in drivers and passengers. The implementation of the cancellation analysis module improves the scientificity and efficiency of order cancellation data processing and promotes the refined management of the platform.

[0038] The responsibility analysis module uses a Gaussian mixture model to determine responsibility, which effectively improves the accuracy and scientificity of the order cancellation responsibility determination. By configuring the preliminary determination rules, the system realizes intelligent discrimination of different cancellation behaviors. When the passenger cancels and the reason is reasonable, the system automatically determines that the driver is not responsible; when the driver cancels and the reason is unreasonable, the system determines that the driver is responsible. The complexity and labor cost of manual review are significantly reduced. The establishment of the Gaussian mixture model enables the system to perform responsibility analysis in multiple dimensions. Through the cross-validation algorithm, the system can randomly initialize the number of Gaussian components and the covariance matrix, thereby realizing a comprehensive evaluation of the driver feedback and customer service return visit results, eliminating the influence of subjective factors. This module not only improves the efficiency of responsibility determination, but also provides a transparent and fair basis for judgment for drivers and passengers, and enhances users' trust in the platform.

[0039] The idle-mileage compensation module is designed to provide a clear and fair compensation mechanism for online ride-hailing drivers. By setting a minimum idle distance of 2 kilometers and a compensation period of one month, the system ensures the rationality and effectiveness of compensation applications. In terms of the deviation between the driver's actual driving route and the preset route, the system can intelligently determine whether the compensation conditions are met to avoid abuse of compensation applications. In terms of the calculation of compensation fees, the system clarifies the standards for basic compensation fees and kilometer compensation fees to ensure the transparency of the compensation amount. Through the visualization platform, drivers can view the compensation calculation results and historical records in real time. This transparent display makes drivers more confident in the compensation mechanism. In addition, the system monitors the driver's compensation application history, intelligently marks abnormal applications, and provides real-time feedback to the driver. This mechanism not only improves the management level of compensation applications, but also promotes a healthy interactive relationship between drivers and the platform. The idle-mileage compensation module effectively improves the driver's job satisfaction and optimizes the platform's operating efficiency.

[0040] The complaint management module provides drivers with an efficient and transparent online complaint channel, enhancing user participation and rights protection. After receiving a notice of no compensation, drivers can conveniently submit the reasons for the complaint and related evidence through the mobile application. The system's automatic recording function ensures the integrity of each complaint information, which is convenient for subsequent review and inquiry. With a 48-hour review time limit, drivers can be informed of the results of the complaint in a timely manner, avoiding dissatisfaction caused by waiting too long. By setting a clear review process and feedback mechanism, the timely handling of complaints and the transparency of the results are ensured. When the complaint is accepted, the system will recalculate the compensation amount and notify the driver in a timely manner, and the complaints that are not accepted will also receive detailed reasons and corresponding terms. This transparent feedback mechanism helps to safeguard the rights and interests of users and enhance drivers' trust in the platform. At the same time, the system sets up an idle compensation board to clearly display the driver's compensation application and complaint status, making the entire process more open and transparent, further improving user satisfaction and loyalty. The complaint management module significantly optimizes the complaint process and improves service quality.

[0041] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A driver idle driving compensation system based on the online car-hailing order cancellation scenario, characterized in that: It includes cancellation analysis module, responsibility analysis module, idle driving compensation module and complaint management module; Cancellation analysis module: collects basic order information and automatically inputs it into the penalty model, divides the order cancellation party into passenger cancellation party and driver cancellation party and marks the cancellation ID, and predefines the cancellation reason options of the cancellation ID according to the cancellation action; Responsibility analysis module: configures a preliminary judgment rule base, builds a Gaussian mixture model, and combines driver feedback and customer service follow-up results to make a probabilistic judgment on the responsibility for order cancellation; Idle driving compensation module: set idle driving compensation standards, define compensation conditions and compensation times, calculate total compensation fees based on basic compensation fees, kilometer compensation fees and waiting compensation fees, and display, monitor and record abnormal applications; Complaint management module: Provide drivers with an online complaint channel through mobile applications, record the reasons and evidence for complaints, set review time limits and feedback results, and transparently display compensation applications and complaint status.

2. According to claim 1, a driver idle driving compensation system based on the online car-hailing order cancellation scenario is characterized by: In the cancellation analysis module, after the order is canceled, the system collects basic information of the order and automatically inputs it into the judgment model. The order canceling party information specified in the order is retrieved through the system log, and the order canceling party is divided into passenger canceling party and driver canceling party and the cancellation ID is marked. The cancellation reason option of the cancellation ID is predefined according to the cancellation action. If the cancellation reason is detected as free text input, NLP technology is used to analyze the input text of the cancellation ID and identify keywords to classify the cancellation reasons.

3. According to claim 1, a driver idle driving compensation system based on the online car-hailing order cancellation scenario is characterized by: In the responsibility analysis module, a preliminary judgment rule library is configured. When the cancellation behavior is caused by the passenger and the cancellation reason is reasonable, the driver is judged to be not responsible. When the cancellation behavior is caused by the driver and the cancellation reason is unreasonable, the driver is judged to be responsible. When it is impossible to judge, the system selects the judgment state, and constructs a Gaussian mixture model to identify the judgment of different dimensions of the judgment state. The number of Gaussian components in the dimension of driver feedback cancellation reasons and the dimension of customer service return visit to passenger results are selected through cross-validation, and each Gaussian component number and covariance matrix are randomly initialized. The responsibility allocation probability of each Gaussian component number corresponding to the dimension of driver feedback cancellation reasons and the dimension of customer service return visit to passenger results is calculated, and the covariance parameter of each Gaussian component number is updated. The judgment of different dimensions of the judgment state is identified through iterative convergence through the log-likelihood function.

4. A driver idle driving compensation system based on the online car-hailing order cancellation scenario according to claim 3, characterized in that: The specific formula of the responsibility allocation probability is: Among them, γ [(i,j)(k,l)] represents the probability of the responsibility distribution corresponding to Gaussian components k and j, respectively, for the driver's feedback cancellation reason dimension i and the customer service return visit passenger result dimension j, N(V i |μ k ,∑k) represents the probability density of the driver’s feedback cancellation reason dimension i under Gaussian distribution k, where V i Indicates the driver's feedback cancellation reason dimension i, μ k represents the mean parameter of the driver feedback cancellation reason dimension i under Gaussian distribution k, ∑k represents the covariance parameter of the driver feedback cancellation reason dimension i under Gaussian distribution k, N(I j |μ l ,∑l) represents the probability density of the customer service return visit result dimension j corresponding to Gaussian distribution j, where I j represents the dimension j of the customer service return visit result, μ l represents the mean parameter of the customer service return visit result dimension j under Gaussian distribution I, ∑l represents the covariance parameter of the customer service return visit result dimension j under Gaussian distribution I, represents the total probability density of all Gaussian components m.

5. According to claim 1, a driver idle driving compensation system based on the online car-hailing order cancellation scenario is characterized by: In the idle driving compensation module, the minimum idle driving distance standard is set at 2 kilometers. The actual distance traveled by the driver from the time the order is accepted to the time the driver cancels the order is recorded and compared with the distances between the starting point and the alighting point set by the passenger. The actual route traveled by the driver is compared with the preset navigation route in the system. If the actual route traveled by the driver deviates from the preset navigation route by 40%, it is judged that the compensation conditions are not met.

6. The driver idle driving compensation system based on the online car-hailing order cancellation scenario according to claim 1 is characterized by: The definition of compensation conditions and compensation times specifically includes: setting the compensation period to one month, limiting the number of idle driving compensation that the driver can apply for within the compensation period to 5 times, the system tracking the number of idle driving compensation that the driver has applied for within the compensation period, and not allowing the driver to submit a new compensation application when the upper limit of the number of idle driving compensation is reached, recording the time when the driver "has been waiting for passengers" and feeding back to the system, and if there is a significant gap between the time fed back by the driver and the estimated arrival time of the navigation, the system feedback is judged to be unreasonable and no compensation is given, connecting to the GPS system to determine whether the driver is close to the boarding point, and within a distance radius of 3m, the system feedback is judged to be reasonable and compensation is given, establishing a compensation rule feedback mechanism and monitoring the driver's compensation application history, and marking abnormal compensation applications.

7. The driver idle driving compensation system based on the online car-hailing order cancellation scenario according to claim 1 is characterized by: The calculation of the total compensation fee based on the basic compensation fee, kilometer compensation fee and waiting compensation fee specifically includes: defining a fixed basic compensation fee to compensate all empty orders that meet the conditions, defining a kilometer compensation fee standard, recording the time the driver waits for passengers to board the vehicle after arriving at the pick-up point and defining a waiting compensation fee, and calculating the total compensation fee based on the basic compensation fee, kilometer compensation fee and waiting compensation fee.

8. The driver idle driving compensation system based on the online car-hailing order cancellation scenario according to claim 1 is characterized by: In the complaint management module, the driver receives a no-compensation notification determined by the system, submits a complaint reason to appeal against the no-compensation result, sets up a driver complaint account through the online complaint channel of the mobile application, adds an idle-mile compensation option to fill in the driver's complaint reason, and notifies the driver to attach driving record evidence including driving route and driving time, passenger conversation record evidence, and emergency evidence including traffic conditions and weather conditions. The system accepts and records all information in the driver's complaint, sets a 48-hour review period to complete the review and feedback of the complaint, and if the complaint is approved, the compensation amount is recalculated and notified to the driver. If the complaint is not approved, the system provides clear reasons and relevant terms, saves all information of each driver complaint to the system complaint log, and sets an idle-mile compensation dashboard to transparently display the driver's compensation application and complaint status, the status of each complaint, the reason for the complaint and the summary of evidence.

Images