Credit evaluation method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a credit assessment method and device, electronic equipment and a storage medium, and is applied to the technical field of data processing, and the method comprises the steps: obtaining historical driving data, historical position data and vehicle state data of a vehicle corresponding to a to-be-assessed user; inputting the historical driving data into the trained driving style prediction model to obtain a predicted target driving style of the to-be-evaluated user; inputting the historical position data into the trained travel stability degree prediction model to obtain a predicted target travel stability degree of the to-be-evaluated user; inputting the vehicle state data into a trained vehicle health assessment model to obtain an assessed target vehicle health degree of the vehicle; and inputting the target driving style, the target travel stability degree and the target vehicle health degree into a trained credit assessment model to obtain an assessed credit score of the to-be-assessed user. The credit evaluation of the user is carried out according to the related historical information of the user vehicle, and the credit evaluation efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of business processing technology, and in particular to a credit assessment method, device, electronic device and storage medium. Background Art

[0002] With the rapid development of society, the Internet financial products of banks have also developed rapidly. While providing credit services on the Internet, banks need to evaluate the credit scores of credit customers in order to reduce the risk of default of customers.

[0003] In the related technology, the existing credit score assessment method is to conduct assessment by having experts design the characteristics of the applicant. The assessment is highly subjective and cannot be replicated in batches, resulting in low accuracy and efficiency of credit score assessment. Therefore, it is urgent to provide a credit assessment method to solve the above technical problems. Summary of the invention

[0004] The main purpose of the embodiments of the present application is to propose a credit assessment method, device, electronic device and storage medium, which can evaluate the credit of a user based on the relevant historical information of the user's vehicle to improve the efficiency of credit assessment.

[0005] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present application proposes a credit assessment method, the method comprising:

[0006] Obtain historical driving data, historical location data, and vehicle status data of the vehicle corresponding to the user to be evaluated;

[0007] Inputting the historical driving data into the trained driving style prediction model to obtain a predicted target driving style of the user to be evaluated;

[0008] Inputting the historical location data into the trained travel stability prediction model to obtain the predicted target travel stability of the user to be evaluated;

[0009] Inputting the vehicle status data into a trained vehicle health assessment model to obtain an assessed target vehicle health level of the vehicle;

[0010] The target driving style, the target travel stability and the target vehicle health are input into the trained credit assessment model to obtain the credit score of the user to be assessed.

[0011] In some embodiments, the target driving style, the target travel stability, and the target vehicle health are input into the trained credit assessment model to obtain the credit score of the user to be assessed, including:

[0012] Encoding the target driving style through an encoder of the trained credit assessment model to obtain a target driving style feature;

[0013] The target travel stability is encoded by an encoder of the trained credit assessment model to obtain a target travel stability feature;

[0014] The target vehicle health level is encoded by an encoder of the trained credit assessment model to obtain a target vehicle health level feature;

[0015] The target driving style feature, the target travel stability feature and the target vehicle health feature are fused through a feature fusion layer of the trained credit assessment model to obtain a fusion feature;

[0016] A target credit score matching the fusion feature is screened out from a plurality of preset credit scores, and the target credit score is determined as the credit score of the user to be evaluated.

[0017] In some embodiments, the feature fusion layer of the trained credit assessment model fuses the target driving style feature, the target travel stability feature, and the target vehicle health feature to obtain a fused feature, including:

[0018] The feature fusion layer of the trained credit assessment model splices the target driving style features, the target travel stability features and the target vehicle health features in a preset splicing order to obtain a fusion feature.

[0019] In some embodiments, selecting a target credit score matching the fusion feature from a plurality of preset credit scores, and determining the target credit score as the credit score of the user to be evaluated, includes:

[0020] Determining the distribution probability of the fusion feature at each preset credit score;

[0021] The preset credit score with the highest probability among the plurality of preset credit scores is determined as the credit score of the user to be evaluated.

[0022] In some embodiments, inputting the historical driving data into a trained driving style prediction model to obtain a predicted target driving style of the user to be evaluated includes:

[0023] encoding the historical driving data through an encoder of the trained driving style prediction model to obtain historical driving features;

[0024] Determine the distribution probability of the historical driving characteristics in each preset driving style;

[0025] The preset driving style with the largest distribution probability is determined as the target driving style of the user to be evaluated.

[0026] In some embodiments, inputting the historical location data into a trained travel stability prediction model to obtain a predicted target travel stability of the user to be evaluated includes:

[0027] The historical location data is encoded by an encoder of the trained travel stability prediction model to obtain historical travel characteristics;

[0028] Determine the distribution probability of the historical travel characteristics at each preset travel stability level;

[0029] The preset travel stability level with the largest distribution probability is determined as the target travel stability level of the user to be evaluated.

[0030] In some embodiments, inputting the vehicle status data into a trained vehicle health assessment model to obtain an assessed target vehicle health level of the vehicle includes:

[0031] The vehicle status data is encoded by an encoder of the trained vehicle health assessment model to obtain a vehicle status feature;

[0032] Determining the distribution probability of the vehicle state characteristic at each preset vehicle health level;

[0033] The preset vehicle health level with the largest distribution probability is determined as the target vehicle health level of the user to be evaluated.

[0034] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application provides a credit assessment device, comprising:

[0035] An acquisition unit, used to acquire historical driving data, historical location data and vehicle status data of a vehicle corresponding to the user to be evaluated;

[0036] A first input unit, configured to input the historical driving data into a trained driving style prediction model to obtain a predicted target driving style of the user to be evaluated;

[0037] A second input unit is used to input the historical location data into the trained travel stability prediction model to obtain the predicted target travel stability of the user to be evaluated;

[0038] A third input unit is used to input the vehicle status data into the trained vehicle health assessment model to obtain an assessed target vehicle health level of the vehicle;

[0039] The fourth input unit is used to input the target driving style, the target travel stability and the target vehicle health into the trained credit assessment model to obtain the credit score of the user to be assessed.

[0040] In some embodiments, the fourth input unit includes:

[0041] A first encoding subunit, configured to encode the target driving style through an encoder of a trained credit assessment model to obtain a target driving style feature;

[0042] A second encoding subunit is used to encode the target travel stability through an encoder of the trained credit assessment model to obtain a target travel stability feature;

[0043] A third encoding subunit is used to encode the health status of the target vehicle through an encoder of the trained credit assessment model to obtain a health status feature of the target vehicle;

[0044] A fusion subunit, configured to fuse the target driving style feature, the target travel stability feature, and the target vehicle health feature through a feature fusion layer of a trained credit assessment model to obtain a fusion feature;

[0045] The screening subunit is used to screen out a target credit score that matches the fusion feature from a plurality of preset credit scores, and determine the target credit score as the credit score of the user to be evaluated.

[0046] In some embodiments, the fusion subunit is used to:

[0047] The feature fusion layer of the trained credit assessment model splices the target driving style features, the target travel stability features and the target vehicle health features in a preset splicing order to obtain a fusion feature.

[0048] In some embodiments, the screening subunit is used to:

[0049] Determining the distribution probability of the fusion feature at each preset credit score;

[0050] The preset credit score with the highest probability among the plurality of preset credit scores is determined as the credit score of the user to be evaluated.

[0051] In some embodiments, the first input unit is used to:

[0052] encoding the historical driving data through an encoder of the trained driving style prediction model to obtain historical driving features;

[0053] Determine the distribution probability of the historical driving characteristics in each preset driving style;

[0054] The preset driving style with the largest distribution probability is determined as the target driving style of the user to be evaluated.

[0055] In some embodiments, the second input unit is used to:

[0056] The historical location data is encoded by an encoder of the trained travel stability prediction model to obtain historical travel characteristics;

[0057] Determine the distribution probability of the historical travel characteristics at each preset travel stability level;

[0058] The preset travel stability level with the largest distribution probability is determined as the target travel stability level of the user to be evaluated.

[0059] In some embodiments, the third input unit is used to:

[0060] The vehicle status data is encoded by an encoder of the trained vehicle health assessment model to obtain a vehicle status feature;

[0061] Determining the distribution probability of the vehicle state characteristic at each preset vehicle health level;

[0062] The preset vehicle health level with the largest distribution probability is determined as the target vehicle health level of the user to be evaluated.

[0063] To achieve the above objectives, a third aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method described in the first aspect is implemented.

[0064] To achieve the above-mentioned purpose, the fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0065] The credit evaluation method, device, electronic device and storage medium proposed in the present application obtain the historical driving data, historical location data and vehicle status data of the vehicle corresponding to the user to be evaluated; input the historical driving data into the trained driving style prediction model to obtain the predicted target driving style of the user to be evaluated; input the historical location data into the trained travel stability prediction model to obtain the predicted target travel stability of the user to be evaluated; input the vehicle status data into the trained vehicle health evaluation model to obtain the evaluated target vehicle health of the vehicle; input the target driving style, the target travel stability and the target vehicle health into the trained credit evaluation model to obtain the evaluated credit score of the user to be evaluated. In this way, compared with the related art that evaluates the credit score by designing the characteristics of the applicant by experts, the present application obtains the historical driving data, historical location data and vehicle status data of the user in daily use of the vehicle, comprehensively evaluates the credit status of the user to be evaluated from different aspects, without designing the characteristics of the applicant, and can more comprehensively understand the user's behavioral characteristics and vehicle status in a multi-dimensional evaluation method, so as to evaluate the user's credit according to the relevant historical information of the user's vehicle, so as to improve the efficiency and comprehensiveness of credit evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 is a flow chart of a credit assessment method provided by an embodiment of the present application;

[0067] Figure 2 yes Figure 1 Flow chart of step S105 in FIG.

[0068] Figure 3 yes Figure 2 Flow chart of step S204 in FIG.

[0069] Figure 4 yes Figure 2 Flow chart of step S205 in FIG.

[0070] Figure 5 yes Figure 1 Flow chart of step S102 in FIG.

[0071] Figure 6 yes Figure 1 Flow chart of step S103 in FIG.

[0072] Figure 7 yes Figure 1 Flow chart of step S104 in FIG.

[0073] Figure 8 is a schematic diagram of the structure of a credit assessment device provided in an embodiment of the present application;

[0074] Fig. 9 It is a schematic diagram of the hardware structure of the electronic device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0075] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0076] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0077] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0078] First, some nouns involved in this application are analyzed:

[0079] Artificial Intelligence (AI): refers to the field of technology and science that allows computers to simulate human intelligence. AI achieves intelligent behavior through:

[0080] 1. Technical basis:

[0081] 1. Algorithms and models:

[0082] Machine learning algorithms: This is one of the core technologies of artificial intelligence. It enables computers to learn patterns and regularities from data without being explicitly programmed. For example, supervised learning algorithms are trained with known input and output data pairs to predict outputs for new data. Unsupervised learning algorithms are used to discover hidden structures and patterns in data.

[0083] Deep learning models: Deep learning is a branch of machine learning that uses multi-layer neural networks to process and analyze data. These neural networks can automatically learn features in data to achieve complex tasks such as image recognition, speech recognition, and natural language processing. For example, convolutional neural networks (CNNs) have achieved great success in the field of image recognition, and recurrent neural networks (RNNs) are suitable for processing sequence data, such as natural language.

[0084] 2. Data:

[0085] A large amount of data is the basis for training artificial intelligence models. Data can come from various sources, such as sensors, databases, the Internet, etc. The quality and quantity of data are crucial to the performance of artificial intelligence models. For example, in image recognition tasks, a large amount of labeled image data is required to train the model so that it can accurately identify different objects.

[0086] Data preprocessing: Before using data to train a model, the data needs to be preprocessed, including data cleaning, data normalization, data enhancement, etc. Data cleaning can remove noise and outliers, data normalization can convert data to a uniform scale, and data enhancement can increase data diversity by randomly transforming the data, thereby improving the generalization ability of the model.

[0087] 2. Intelligent Expression:

[0088] 1. Perception ability:

[0089] Computer Vision: Enables computers to understand and analyze image and video data. For example, face recognition technology can identify faces in images and extract facial feature information. Object detection technology can identify different objects in images and determine their location and category.

[0090] Speech recognition: enables computers to understand and process human speech signals. For example, voice assistants can understand user instructions through speech recognition technology and perform corresponding operations. Speech synthesis technology can convert text into speech to achieve human-computer interaction.

[0091] Natural Language Processing: enables computers to understand, generate, and process human language. For example, machine translation can translate from one language to another. Text classification can classify text data into different categories. Sentiment analysis can analyze the emotional tendencies in text, such as positive, negative, or neutral.

[0092] 2. Learning and reasoning ability:

[0093] Autonomous learning: AI systems can improve their performance by continuously learning from new data. For example, online learning algorithms can update models in real time as new data arrives to adapt to changing circumstances.

[0094] Reasoning and decision-making: AI systems can reason and make decisions based on existing knowledge and data. For example, expert systems can reason with rules and facts in a knowledge base to solve problems in a specific field. Reinforcement learning algorithms can learn optimal decision-making strategies through interaction with the environment to achieve specific goals.

[0095] 3. Creativity and adaptability:

[0096] Creative thinking: Some AI systems can demonstrate some creativity, such as generating works of art, music, or literature. These systems typically use deep learning techniques to generate new works by learning from a large number of existing works.

[0097] Adaptability: AI systems can adapt to different environments and tasks. For example, a robot can adjust its behavior to complete different tasks by sensing the environment and learning tasks.

[0098] Based on this, the embodiments of the present application provide a credit assessment method, device, electronic device and storage medium, which are intended to assess the credit of a user based on relevant historical information of the user's vehicle to improve the efficiency of credit assessment.

[0099] The credit assessment method, device, electronic device and storage medium provided in the embodiments of the present application are specifically described through the following embodiments. First, the credit assessment method in the embodiments of the present application is described.

[0100] The credit assessment method provided in the embodiment of the present application relates to the field of data processing technology. The credit assessment method provided in the embodiment of the present application can be applied to a terminal, can be applied to a server side, or can be software running in a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the credit assessment method, etc., but is not limited to the above forms.

[0101] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0102] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0103] Figure 1 is an optional flow chart of the credit assessment method provided in the embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S106.

[0104] Step S101, obtaining historical driving data, historical location data and vehicle status data of the vehicle corresponding to the user to be evaluated.

[0105] Among them, historical driving data is the driving related information of the vehicle in the past period of time. It includes data such as the speed change, acceleration, braking frequency, mileage, etc. of the vehicle, which can reflect the driving dynamics of the vehicle and the operating habits of the driver. These data can be collected through the on-board diagnostic system (OBD), the built-in sensors of the vehicle, or special vehicle data acquisition equipment. These devices can record the driving status of the vehicle in real time and store the data for subsequent analysis. After the data is collected, preliminary sorting and screening may be required to remove outliers and erroneous data. Historical location data is the geographical location information of the vehicle at different time points. It is usually obtained through the global positioning system (GPS), including latitude and longitude coordinates, driving trajectory, and stopover locations. These data can be used to analyze the user's travel patterns and frequented places. The location data can be obtained by using the GPS device on the vehicle or a mobile device connected to the vehicle (such as a smartphone). The data can be recorded at a certain time interval, such as recording the location information once every minute or every hour. The collected data can be visualized through map software or data analysis tools to better understand the user's travel stability. The vehicle status data reflects the current mechanical condition and performance indicators of the vehicle. For example, the vehicle's engine status, tire pressure, fuel consumption, fault codes, etc. This data can help assess the health and reliability of the vehicle. Vehicle status data can be obtained through on-board sensors, the vehicle's electronic control unit (ECU), or specialized vehicle testing equipment. Some vehicle manufacturers also provide remote diagnostic services that can transmit vehicle status data to a server for analysis via the Internet. The collected data can be compared with the vehicle's technical specifications and standards to determine the vehicle's health status.

[0106] Specifically, after collecting the historical location data, the multiple location data can be divided according to time periods, and the first historical location data for the morning and evening peak commuting period and the second historical location data for the non-morning and evening peak commuting period are divided. The first historical location data for the morning and evening peak commuting period describes the user's commuting stability to reflect whether the user has a stable job, and the second historical location data for the non-morning and evening peak commuting period describes the user's spending power in the places he or she goes to in his or her spare time to reflect the user's spending power. In addition, since the user may not stop at the same location every time, even if the user's final destination is the same, the location data will be different. Therefore, for each historical location information, a cluster analysis can be performed on the multiple historical location information to cluster the multiple historical location information in the same area, so as to reduce the amount of calculation in the subsequent calculation process.

[0107] Step S102: inputting the historical driving data into the trained driving style prediction model to obtain a predicted target driving style of the user to be evaluated.

[0108] The driving style prediction model is a trained machine learning model that can infer the user's driving style based on the vehicle's historical driving data. Driving styles can be divided into many types, such as aggressive, stable, and economical. Different driving styles reflect the user's behavioral characteristics and risk preferences during driving. The target driving style is the driving style type of the user to be evaluated predicted by the driving style prediction model. For example, if the model predicts that the user's driving style is stable, then the target driving style is "stable".

[0109] Specifically, the training method of the driving style prediction model can be: prepare a large amount of historical driving data with labeled driving styles as a training set. Then, select a suitable machine learning algorithm, such as decision tree, random forest, neural network, etc., to train the training set. During the training process, the model will learn the relationship between different driving styles and driving data. After the training is completed, the historical driving data of the user to be evaluated is input into the model, and the model will output the predicted driving style.

[0110] Step S103: input the historical location data into the trained travel stability prediction model to obtain the predicted target travel stability of the user to be evaluated.

[0111] The travel stability prediction model is a machine learning model used to evaluate the user's travel stability based on the historical location data of the vehicle. The travel stability can reflect the user's life pattern and work stability, such as whether the user often travels at a fixed time and place, whether the travel trajectory has a certain degree of repetitiveness, etc. The target travel stability is the travel stability index of the user to be evaluated predicted by the model. It can be expressed as a numerical value, such as a range from 0 to 1, and the higher the value, the higher the travel stability.

[0112] Specifically, the output result of the travel stability prediction model can be a specific value or a classification result, such as three levels of travel stability: high, medium, and low. According to specific needs, the output result can be further processed and interpreted to better understand the user's travel stability.

[0113] The training method of the stability prediction model is similar to that of the driving style prediction model. It is necessary to prepare historical location data with labeled travel stability as a training set. Select a suitable algorithm for training so that the model can learn the relationship between travel stability and location data. Input the historical location data of the user to be evaluated into the model, and the model will output the predicted travel stability.

[0114] Step S104: input the vehicle status data into the trained vehicle health assessment model to obtain an assessed target vehicle health level of the vehicle.

[0115] The vehicle health assessment model is a machine learning model used to assess the health status of a vehicle. By analyzing the vehicle status data, it is possible to determine whether the various components of the vehicle are working properly and whether there is a potential risk of failure. The target vehicle health level is a vehicle health status indicator obtained by the vehicle health assessment model. It can be expressed as a numerical value, such as a score from 0 to 100, where a higher score indicates a better vehicle health status.

[0116] Specifically, the model output can be a specific value or a classification result, such as three levels of vehicle health: good, average, and poor. Based on the evaluation results, vehicle maintenance suggestions can be provided to users or used as a reference factor for credit evaluation.

[0117] The vehicle health assessment model is trained by using vehicle status data with labeled vehicle health as a training set and selecting a suitable algorithm for training. The model can learn the relationship between different vehicle status and health. The status data of the vehicle to be assessed is input into the model, and the model will output the vehicle health assessment result.

[0118] Step S105, inputting the target driving style, the target travel stability and the target vehicle health into the trained credit assessment model to obtain the credit score of the user to be assessed.

[0119] Among them, the credit assessment model is a machine learning model used to assess the user's credit status. The model comprehensively considers multiple factors such as the user's driving style, travel stability, and vehicle health to predict the user's credit risk and credit rating. The credit score is the user's credit score obtained by the credit assessment model. The credit score can reflect the user's credit status and risk level, and is usually used to decide whether to grant the user financial services such as loans and insurance, as well as to determine the conditions and interest rates of the services.

[0120] Specifically, credit scores can be interpreted and applied according to specific business needs. For example, different credit rating intervals can be set, and users can be divided into different levels according to their credit scores. Financial institutions can decide whether to approve loan applications, determine loan amounts and interest rates, etc. based on users' credit ratings. At the same time, credit scores can also be dynamically adjusted as user behavior and vehicle conditions change, to achieve real-time credit assessment and risk management.

[0121] The credit assessment model is trained using sample data with labeled credit scores. The sample data should contain features such as target driving style, target travel stability, target vehicle health, and corresponding credit scores. Select a suitable algorithm for training so that the model can learn the relationship between these features and credit scores. Input the target driving style, target travel stability, and target vehicle health of the user to be assessed into the model, and the model will output a credit score.

[0122] Steps S101 to S105 shown in the embodiment of the present application are performed by obtaining historical driving data, historical location data and vehicle status data of the vehicle corresponding to the user to be evaluated; inputting the historical driving data into the trained driving style prediction model to obtain the predicted target driving style of the user to be evaluated; inputting the historical location data into the trained travel stability prediction model to obtain the predicted target travel stability of the user to be evaluated; inputting the vehicle status data into the trained vehicle health assessment model to obtain the target vehicle health of the assessed vehicle; inputting the target driving style, the target travel stability and the target vehicle health into the trained credit assessment model to obtain the assessed credit score of the user to be evaluated. Therefore, compared with the related art in which experts design the characteristics of the applicant to evaluate the credit score, this application obtains the historical driving data, historical location data and vehicle status data of the user in daily use of the vehicle to comprehensively evaluate the credit status of the user to be evaluated from different aspects. There is no need to design the characteristics of the applicant. The multi-dimensional evaluation method can more comprehensively understand the user's behavioral characteristics and vehicle status, and realize the evaluation of the user's credit according to the relevant historical information of the user's vehicle, so as to improve the efficiency and comprehensiveness of credit evaluation.

[0123] See also Figure 2 In some embodiments, step S105 may include but is not limited to steps S201 to S205:

[0124] Step S201, encoding the target driving style through an encoder of a trained credit assessment model to obtain target driving style features.

[0125] The encoder is a part of the credit assessment model, and its main function is to convert the input target driving style information into a specific feature representation. This feature representation can be better understood and processed by the model, and can also capture the key features of the target driving style.

[0126] Specifically, the encoder first takes the target driving style as the input of the encoder. The target driving style can be several pre-defined types, such as aggressive, stable, economical, etc., or a continuous numerical representation predicted by other models.

[0127] The encoder is usually composed of multiple layers of neural networks, such as fully connected layers, convolutional layers, or recurrent neural network layers. The input target driving style data is processed by these neural network layers to gradually extract higher-level features.

[0128] In each layer, data is extracted and transformed through linear transformation and nonlinear activation function. For example, the fully connected layer can perform linear transformation on the input data through matrix multiplication and addition of bias terms, and then perform nonlinear activation through activation function (such as ReLU function) to enhance the expressive power of the model.

[0129] After being processed by a multi-layer neural network, the target driving style feature is finally obtained. This feature is a vector, and its dimension and specific value depend on the structure and training process of the encoder. The target driving style feature can better reflect the essential characteristics of the target driving style and provide more valuable information for subsequent credit assessment.

[0130] Step S202: Encode the target travel stability through the encoder of the trained credit assessment model to obtain the target travel stability characteristics.

[0131] The target travel stability level is represented by a numerical value, such as a continuous value from 0 to 1, or divided into several discrete levels such as high, medium, and low.

[0132] The encoder processes the input target travel stability data. Similarly, the encoder can be composed of a multi-layer neural network, which gradually extracts the characteristics of the target travel stability through linear transformation and nonlinear activation function. Finally, the target travel stability characteristics are obtained. This feature vector can better reflect the characteristics of the target travel stability and provide a basis for credit assessment.

[0133] Step S203, encoding the target vehicle health status through the encoder of the trained credit assessment model to obtain the target vehicle health status characteristics.

[0134] The encoder encodes the target vehicle health data. The structure and training process of the encoder are similar to those of the encoders for the target driving style and target travel stability. The characteristics of the target vehicle health are extracted through the processing of a multi-layer neural network.

[0135] Specifically, after obtaining the target vehicle health feature vector, this vector can reflect the key information of the vehicle health and provide an important reference for credit assessment.

[0136] Step S204, through the feature fusion layer of the trained credit assessment model, the target driving style feature, the target travel stability feature and the target vehicle health feature are fused to obtain a fused feature.

[0137] The purpose of the trained feature fusion layer is to integrate features from different sources (target driving style features, target travel stability features, and target vehicle health features) to obtain a more comprehensive and representative fusion feature. This can make full use of the complementary information between different features and improve the accuracy of credit assessment.

[0138] Step S205, selecting a target credit score that matches the fusion feature from a plurality of preset credit scores, and determining the target credit score as the credit score of the user to be evaluated.

[0139] The preset credit score can be set according to business needs and experience. For example, the credit score can be divided into several intervals, such as high credit score (80-100), medium credit score (60-79), low credit score (40-59) and poor credit score (0-39). Each interval corresponds to a different credit rating and risk level.

[0140] Specifically, the fused features are input into the last layer of the credit assessment model, which is usually a classification layer or a regression layer. If it is a classification problem, the softmax function can be used to map the fused features to the probability distribution between different credit partitions; if it is a regression problem, linear regression or neural network methods can be used to directly predict a continuous credit score.

[0141] According to the predicted probability distribution or continuous credit score, the best matching target credit score is selected from the preset credit scores. For example, if the predicted probability distribution shows that the user has the highest probability of belonging to the high credit zone, a specific score (such as 85 points) in the high credit zone can be used as the target credit score.

[0142] The target credit score is determined as the credit score of the user to be evaluated. This credit score can be used to decide whether to grant the user financial services such as loans and insurance, as well as to determine the conditions of the services.

[0143] In this way, by encoding the target driving style, target travel stability and target vehicle health, we can dig deeper into information related to user credit from different angles. Driving style reflects the user's behavior habits and risk preferences, travel stability reflects the regularity of life and work and the user's consumption ability, and vehicle health is related to the value and stability of assets. These multi-dimensional features can provide a more comprehensive perspective for credit assessment and avoid the limitations of single-factor assessment.

[0144] See also Figure 3 In some embodiments, step S204 may include but is not limited to step S301:

[0145] Step S301, splicing the target driving style feature, the target travel stability feature and the target vehicle health feature according to a preset splicing order through the feature fusion layer of the trained credit assessment model to obtain a fusion feature.

[0146] Among them, the order of splicing the target driving style features, target travel stability features, and target vehicle health features is determined according to business needs and data characteristics. For example, the target driving style features can be spliced ​​first, then the target travel stability features, and finally the target vehicle health features; or the splicing order can be determined according to the importance of different features to credit assessment. Considering that the dimensions and data types of different features may be different, it is necessary to ensure that the splicing order can make the fused features have a reasonable structure and an easy-to-process form.

[0147] Specifically, the target driving style features, target travel stability features, and target vehicle health features are obtained from the encoder output of the credit assessment model. These feature vectors are obtained after the encoder processes the corresponding raw data and have specific dimensions and numerical representations. Ensure the accuracy and completeness of the feature vectors to avoid data loss or errors.

[0148] According to the preset splicing order, the target driving style feature, the target travel stability feature and the target vehicle health feature are spliced ​​in sequence. Specifically, these feature vectors can be connected in corresponding dimensions to form a longer feature vector.

[0149] For example, if the target driving style feature is a vector with a length of n1, the target travel stability feature is a vector with a length of n2, and the target vehicle health feature is a vector with a length of n3, after splicing in the preset splicing order, the length of the obtained fused feature vector is n1+n2+n3.

[0150] The concatenated fused feature vector is used as the output of the feature fusion layer and provided to the subsequent parts of the credit assessment model for processing. This fused feature vector integrates information from multiple aspects and provides a more comprehensive feature representation for credit assessment.

[0151] In this way, determining the splicing order based on data characteristics can make full use of the potential information in the data. For example, if the data shows that a certain feature has a stronger correlation with the credit assessment result, then the feature can be placed at the front of the splicing order to highlight its importance. This can improve the accuracy and practicality of the model.

[0152] See also Figure 4 In some embodiments, step S205 may include but is not limited to steps S401 to S402:

[0153] Step S401, determining the distribution probability of the fusion feature in each preset credit score.

[0154] Among them, classification algorithms in machine learning, such as Bayesian classifiers, support vector machines (SVM), logistic regression, etc., or neural network classifiers in deep learning are used to determine the distribution probability of the fusion features in each preset credit score.

[0155] Collect a batch of existing user data, which includes the user's fused features and the corresponding known credit scores. Divide this data into a training set and a test set. Use the training set data to train the selected classification algorithm. During the training process, the model learns the relationship between fused features and different credit scores so that it can predict the probability of new fused features at each preset credit score. Input the fused features of the user to be evaluated into the trained model, and the model outputs the distribution probability of the fused features at each preset credit score.

[0156] Step S402: Determine the preset credit score with the highest probability among the plurality of preset credit scores as the credit score of the user to be evaluated.

[0157] The probability of each preset credit score of the fusion features output by the model is compared. The preset credit score with the highest probability is selected as the credit score of the user to be evaluated. This credit score represents the most likely evaluation result of the model on the user's credit status.

[0158] For example, suppose there are four preset credit score intervals of 60-70, 71-80, 81-90, and 91-100. The model calculates that the probability of the fusion features of the user to be evaluated in these four intervals is 0.2, 0.3, 0.4, and 0.1 respectively. Then, the preset credit score interval of 81-90 with the highest probability is determined as the credit score interval of the user. If a more specific credit score is required, a representative value in this interval, such as 85 points, can be taken as the user's final credit score.

[0159] In this way, by determining the distribution probability of the fusion feature in each preset credit score, it is possible to determine which range the user's credit status is most likely to fall into. This provides financial institutions with a preliminary credit assessment range, which helps to quickly understand the user's credit level, such as whether it is in a higher credit range or a lower credit range, so as to take appropriate decision-making measures. Determine the specific credit score: After determining the credit score range with the highest probability, further take a representative value as the final credit score, making the credit assessment result more specific and intuitive. This helps financial institutions have a clearer reference basis when formulating specific financial service conditions such as loan amounts and interest rates, and also makes it easier for users to understand their credit status.

[0160] See also Figure 5 In some embodiments, step S102 may include but is not limited to steps S501 to S503:

[0161] Step S501 : encoding the historical driving data through an encoder of the trained driving style prediction model to obtain historical driving characteristics.

[0162] The encoder can be composed of a multi-layer neural network, such as a fully connected layer, a convolutional layer, or a recurrent neural network layer, etc. The appropriate encoder structure is selected according to the characteristics of the historical driving data and the requirements of the model.

[0163] Input historical driving data: The historical driving data of the user to be evaluated is used as the input of the encoder. This data can be obtained from vehicle sensors, vehicle data recording devices, etc.

[0164] Encoding: The historical driving data is processed by the various layers of neural networks of the encoder. In each layer, the data is subjected to linear transformation and nonlinear activation function to extract and transform features. For example, the fully connected layer can perform linear transformation on the input data through matrix multiplication and addition of bias terms, and then perform nonlinear activation through the activation function to enhance the expression ability of the model. After being processed by the multi-layer neural network, the historical driving features are obtained.

[0165] Step S502: determining the distribution probability of the historical driving characteristics in each preset driving style.

[0166] The historical driving characteristics of the user to be evaluated are input into the trained model, and the model outputs the distribution probability of the historical driving characteristics in each preset driving style.

[0167] Step S503: determining the preset driving style with the largest distribution probability as the target driving style of the user to be evaluated.

[0168] The historical driving characteristics output by the model are compared in the probability of each preset driving style. The preset driving style with the highest probability is selected as the target driving style of the user to be evaluated. This driving style represents the most likely evaluation result of the model on the user's driving behavior.

[0169] Specific processing method: The result of the model output is usually a probability distribution, which indicates the possibility of the user belonging to different driving styles. The driving style with the highest probability can be selected as the target driving style, or the user's driving style can be determined based on a certain threshold. For example, if the probability that the user belongs to the stable driving style is more than 50%, the target driving style is determined to be stable.

[0170] In this way, by inputting historical driving characteristics into the trained model, the distribution probability of historical driving characteristics in each preset driving style can be obtained. This data- and model-based approach avoids the uncertainty of subjective judgment and provides a scientific and quantitative way to evaluate the possibility of a user's driving style. Through probability distribution, financial institutions or other relevant parties can have a more comprehensive understanding of the user's driving style tendencies, rather than just determining a single driving style category. This helps to more accurately assess the user's risk level and credit status, providing richer information for decision-making.

[0171] See also Figure 6 In some embodiments, step S103 may include but is not limited to steps S601 to S602:

[0172] Step S601, encoding the historical location data through the encoder of the trained travel stability prediction model to obtain historical travel characteristics.

[0173] The historical location data of the user to be evaluated is used as the input of the encoder. This data can be obtained through GPS devices on the vehicle, mobile device positioning, etc. The historical location data is processed by each layer of the encoder's neural network. In each layer, the data is extracted and transformed through the action of linear transformation and nonlinear activation function. For example, the fully connected layer can perform linear combination and nonlinear activation on the input data to extract higher-level features. After being processed by the multi-layer neural network, the historical travel features are obtained.

[0174] Step S602: determining the distribution probability of the historical travel characteristics at each preset travel stability level.

[0175] The historical travel characteristics of the user to be evaluated are input into the trained model, and the model outputs the distribution probability of the historical travel characteristics at each preset travel stability level.

[0176] Step S603: determine the preset travel stability level with the largest distribution probability as the target travel stability level of the user to be evaluated.

[0177] The historical travel characteristics output by the model are compared at the probability of each preset travel stability level. The preset travel stability level with the highest probability is selected as the target travel stability level of the user to be evaluated. This travel stability level represents the most likely evaluation result of the model on the user's travel regularity and stability.

[0178] In this way, the historical travel characteristics are input into the trained model to obtain the distribution probability of the historical travel characteristics at each preset travel stability level. This data- and model-based method avoids the uncertainty of subjective judgment and provides a scientific and quantitative way to evaluate the possibility of user travel stability. Through probability distribution, we can fully understand the various possibilities of user travel stability, rather than simply determining a single stability category.

[0179] See also Figure 7 In some embodiments, step S104 may also include but is not limited to steps S701 to S703:

[0180] Step S701, encoding the vehicle status data through the encoder of the trained vehicle health assessment model to obtain vehicle status characteristics.

[0181] The vehicle status data of the user to be evaluated is used as the input of the encoder. This data can be obtained through vehicle sensors, vehicle diagnostic systems, etc. The vehicle status data is processed by each layer of the neural network of the encoder. In each layer, the data is extracted and converted through the action of linear transformation and nonlinear activation function. For example, the fully connected layer can perform linear combination and nonlinear activation on the input data to extract higher-level features. After being processed by the multi-layer neural network, the vehicle status features are obtained.

[0182] Step S702, determining the distribution probability of the vehicle state feature at each preset vehicle health level.

[0183] The vehicle status characteristics of the user to be evaluated are input into the trained model, and the model outputs the distribution probability of the vehicle status characteristics at each preset vehicle health level.

[0184] Step S703: determine the preset vehicle health level with the largest distribution probability as the target vehicle health level of the user to be evaluated.

[0185] The vehicle status features output by the model are compared with the probability of each preset vehicle health level. The preset vehicle health level with the highest probability is selected as the target vehicle health level of the user to be evaluated. This health level represents the most likely evaluation result of the model on the vehicle health status.

[0186] In this way, the probability of vehicle status characteristics at multiple preset vehicle health levels can be determined, which can fully consider the diversity and uncertainty of vehicle status. The health status of a vehicle may be affected by many factors, such as the use environment, maintenance conditions, driving habits, etc. The probability distribution can reflect this complexity. This can make the evaluation results more objective and accurate, and provide richer information for decision-making.

[0187] Specifically, after determining the credit score of the user to be evaluated, the loan amount of the user can be calculated based on the credit score. And the following contents are implemented: 1. Real-time monitoring and analysis: Use the Internet of Vehicles platform to collect and analyze vehicle driving data, location information and user behavior data in real time. Through preset algorithms and rules, key indicators are monitored, such as the number of emergency braking, speeding ratio, abnormal increase in night driving frequency, or frequent appearance of vehicles in high-risk areas. 2. Threshold setting: According to historical data and business experience, reasonable thresholds or ranges are set for key indicators. When a certain indicator is monitored to exceed the threshold or range, the system automatically triggers a risk warning. 3. Warning notification: After the risk warning is triggered, the loan user, loan institution and relevant risk control personnel are immediately notified through SMS, email, APP push, etc. The specific content of the abnormal behavior or risk signal should be clearly stated in the warning notification, and immediate measures should be taken to verify or deal with it. II. Evaluation and adjustment measures of actual situation 1. Abnormal situation verification: After receiving the warning notification, the risk control personnel shall immediately verify the abnormal situation. Understand the actual situation through telephone follow-up, on-site investigation, etc., and confirm whether there is risk behavior. 2. Adjustment of loan conditions: If risky behavior is verified, but the user still has the ability to repay, the lending institution may adjust the loan conditions according to the actual situation. For example, increase the loan interest rate, shorten the loan period, increase guarantee measures, etc. to reduce the loan risk. 3. Take corresponding measures: For serious default or fraudulent behavior, the lending institution should immediately take corresponding measures to protect its own rights and interests. Including but not limited to suspending loan issuance, early loan recovery, etc. 4. Continuous monitoring and adjustment: After taking corresponding measures, the lending institution needs to continuously monitor the user's repayment situation and vehicle status. Adjust the risk control strategy and loan conditions in a timely manner according to the actual situation to ensure the safety and sound operation of the loan assets. III. Risk management and optimization 1. Risk database construction: Establish a risk database to record the characteristics and processing methods of various abnormal behaviors and risk signals. Improve the accuracy and timeliness of risk warnings by continuously accumulating and optimizing risk data. 2. Iteration of risk control model: Continuously optimize the risk control model based on the feedback data in the risk monitoring and adjustment process. Improve the predictive ability and adaptability of the risk control model by introducing new algorithms and rules. 3. Internal training and communication: Conduct internal training for risk control personnel regularly to improve their ability to identify risk signals and abnormal behaviors. Strengthen communication with loan users, clarify loan conditions and management requirements, and reduce risks caused by information asymmetry.

[0188] See also Figure 8 The present application also provides a credit evaluation device that can implement the above credit evaluation method. The device includes:

[0189] An acquisition unit, used to acquire historical driving data, historical location data and vehicle status data of a vehicle corresponding to the user to be evaluated;

[0190] A first input unit, configured to input the historical driving data into a trained driving style prediction model to obtain a predicted target driving style of the user to be evaluated;

[0191] A second input unit is used to input the historical location data into the trained travel stability prediction model to obtain the predicted target travel stability of the user to be evaluated;

[0192] A third input unit is used to input the vehicle status data into the trained vehicle health assessment model to obtain an assessed target vehicle health level of the vehicle;

[0193] The fourth input unit is used to input the target driving style, the target travel stability and the target vehicle health into the trained credit assessment model to obtain the credit score of the user to be assessed.

[0194] In some embodiments, the fourth input unit includes:

[0195] A first encoding subunit, configured to encode the target driving style through an encoder of a trained credit assessment model to obtain a target driving style feature;

[0196] A second encoding subunit is used to encode the target travel stability through an encoder of the trained credit assessment model to obtain a target travel stability feature;

[0197] A third encoding subunit is used to encode the health status of the target vehicle through an encoder of the trained credit assessment model to obtain a health status feature of the target vehicle;

[0198] A fusion subunit, configured to fuse the target driving style feature, the target travel stability feature, and the target vehicle health feature through a feature fusion layer of a trained credit assessment model to obtain a fusion feature;

[0199] The screening subunit is used to screen out a target credit score that matches the fusion feature from a plurality of preset credit scores, and determine the target credit score as the credit score of the user to be evaluated.

[0200] In some embodiments, the fusion subunit is used to:

[0201] The feature fusion layer of the trained credit assessment model splices the target driving style features, the target travel stability features and the target vehicle health features in a preset splicing order to obtain a fusion feature.

[0202] In some embodiments, the screening subunit is used to:

[0203] Determining the distribution probability of the fusion feature at each preset credit score;

[0204] The preset credit score with the highest probability among the plurality of preset credit scores is determined as the credit score of the user to be evaluated.

[0205] In some embodiments, the first input unit is used to:

[0206] encoding the historical driving data through an encoder of the trained driving style prediction model to obtain historical driving features;

[0207] Determine the distribution probability of the historical driving characteristics in each preset driving style;

[0208] The preset driving style with the largest distribution probability is determined as the target driving style of the user to be evaluated.

[0209] In some embodiments, the second input unit is used to:

[0210] The historical location data is encoded by an encoder of the trained travel stability prediction model to obtain historical travel characteristics;

[0211] Determine the distribution probability of the historical travel characteristics at each preset travel stability level;

[0212] The preset travel stability level with the largest distribution probability is determined as the target travel stability level of the user to be evaluated.

[0213] In some embodiments, the third input unit is used to:

[0214] The vehicle status data is encoded by an encoder of the trained vehicle health assessment model to obtain a vehicle status feature;

[0215] Determining the distribution probability of the vehicle state characteristic at each preset vehicle health level;

[0216] The preset vehicle health level with the largest distribution probability is determined as the target vehicle health level of the user to be evaluated.

[0217] The credit assessment device provided in the embodiment of the present application obtains historical driving data, historical position data and vehicle status data of the vehicle corresponding to the user to be assessed through an acquisition unit; the first input unit inputs the historical driving data into the trained driving style prediction model to obtain the predicted target driving style of the user to be assessed; the second input unit inputs the historical position data into the trained travel stability prediction model to obtain the predicted target travel stability of the user to be assessed; the third input unit inputs the vehicle status data into the trained vehicle health assessment model to obtain the assessed target vehicle health; the fourth input unit inputs the target driving style, the target travel stability and the target vehicle health into the trained credit assessment model to obtain the assessed credit score of the user to be assessed. In this way, compared with the related art that evaluates credit scores by designing the characteristics of applicants by experts, this application obtains the historical driving data, historical location data and vehicle status data of users in daily use of vehicles, and comprehensively evaluates the credit status of users to be evaluated from different aspects. There is no need to design the characteristics of applicants. The multi-dimensional evaluation method can more comprehensively understand the user's behavioral characteristics and vehicle conditions, and evaluate the user's credit according to the relevant historical information of the user's vehicle, so as to improve the efficiency and comprehensiveness of credit evaluation. By encoding the target driving style, target travel stability and target vehicle health respectively, it is possible to deeply explore information related to user credit from different angles. Driving style reflects the user's behavioral habits and risk preferences, travel stability reflects the regularity of life and work and the user's consumption ability, and vehicle health is related to the value and stability of assets. These multi-dimensional features can provide a more comprehensive perspective for credit evaluation and avoid the limitations of single factor evaluation. Determining the splicing order based on data characteristics can make full use of the potential information in the data. For example, if the data shows that a certain feature has a stronger correlation with the credit evaluation result, then the feature can be placed in front of the splicing order to highlight its importance. This can improve the accuracy and practicality of the model. By determining the distribution probability of the fused features in each preset credit score, it is possible to determine which range the user's credit status is most likely to fall into. This provides financial institutions with a preliminary credit assessment range, which helps to quickly understand the user's credit level, such as whether it is in a higher credit range or a lower credit range, so as to take appropriate decision-making measures. Determine the specific credit score: After determining the credit score range with the highest probability, further take a representative value as the final credit score, making the credit assessment results more specific and intuitive. This helps financial institutions have a clearer reference basis when formulating specific financial service conditions such as loan amounts and interest rates, and also makes it easier for users to understand their credit status. By inputting historical driving features into the trained model, the distribution probability of historical driving features in each preset driving style can be obtained.This data- and model-based approach avoids the uncertainty of subjective judgment and provides a scientific and quantitative way to evaluate the possibility of a user's driving style. Through probability distribution, financial institutions or other relevant parties can have a more comprehensive understanding of the user's driving style tendencies, rather than simply determining a single driving style category. This helps to more accurately assess the user's risk level and credit status, providing richer information for decision-making. The historical travel characteristics are input into the trained model to obtain the distribution probability of the historical travel characteristics at each preset travel stability level. This data- and model-based approach avoids the uncertainty of subjective judgment and provides a scientific and quantitative way to evaluate the possibility of a user's travel stability level. Through probability distribution, the various possibilities of the user's travel stability level can be fully understood, rather than simply determining a single stability level category. Determining the probability of vehicle status characteristics at multiple preset vehicle health levels can fully consider the diversity and uncertainty of vehicle status. The health status of a vehicle may be affected by many factors, such as the use environment, maintenance conditions, driving habits, etc., and the probability distribution can reflect this complexity. This can make the evaluation results more objective and accurate, providing richer information for decision-making.

[0218] The specific implementation of the credit evaluation device is basically the same as the specific implementation of the above-mentioned credit evaluation method, and will not be repeated here.

[0219] The embodiment of the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above credit assessment method when executing the computer program. The electronic device can be any smart terminal including a tablet computer, a car computer, etc.

[0220] See also Fig. 9 , Fig. 9 The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes:

[0221] The processor 901 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0222] The memory 902 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 902 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 902, and the processor 901 calls and executes the credit assessment method of the embodiment of this application;

[0223] Input / output interface 903, used to implement information input and output;

[0224] Communication interface 904, used to realize communication interaction between the device and other devices, which can be realized by wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.);

[0225] A bus 905 that transmits information between various components of the device (e.g., the processor 901, the memory 902, the input / output interface 903, and the communication interface 904);

[0226] The processor 901 , the memory 902 , the input / output interface 903 and the communication interface 904 are connected to each other in communication within the device via a bus 905 .

[0227] An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program, and the computer program implements the above-mentioned credit assessment method when executed by a processor.

[0228] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0229] The credit assessment method, credit assessment device, electronic device and storage medium provided in the embodiments of the present application obtain historical driving data, historical location data and vehicle status data of the vehicle corresponding to the user to be assessed; input the historical driving data into a trained driving style prediction model to obtain a predicted target driving style of the user to be assessed; input the historical location data into a trained travel stability prediction model to obtain a predicted target travel stability of the user to be assessed; input the vehicle status data into a trained vehicle health assessment model to obtain a target vehicle health of the assessed vehicle; input the target driving style, the target travel stability and the target vehicle health into the trained credit assessment model to obtain an assessed credit score of the user to be assessed. Therefore, compared with the related art in which experts design the characteristics of the applicant to evaluate the credit score, this application obtains the historical driving data, historical location data and vehicle status data of the user in daily use of the vehicle to comprehensively evaluate the credit status of the user to be evaluated from different aspects. There is no need to design the characteristics of the applicant. The multi-dimensional evaluation method can more comprehensively understand the user's behavioral characteristics and vehicle status, and realize the evaluation of the user's credit according to the relevant historical information of the user's vehicle, so as to improve the efficiency and comprehensiveness of credit evaluation.

[0230] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0231] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0232] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0233] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0234] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0235] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0236] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0237] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0238] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0239] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0240] The preferred embodiments of the present application are described above with reference to the accompanying drawings, but the scope of the rights of the present application is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present application should be within the scope of the rights of the present application.

Claims

1. A credit assessment method, characterized in that: include: Obtain historical driving data, historical location data, and vehicle status data of the vehicle corresponding to the user to be evaluated; Inputting the historical driving data into the trained driving style prediction model to obtain a predicted target driving style of the user to be evaluated; Inputting the historical location data into the trained travel stability prediction model to obtain the predicted target travel stability of the user to be evaluated; Inputting the vehicle status data into a trained vehicle health assessment model to obtain an assessed target vehicle health level of the vehicle; The target driving style, the target travel stability and the target vehicle health are input into the trained credit assessment model to obtain the credit score of the user to be assessed.

2. The credit assessment method according to claim 1, characterized in that: The target driving style, the target travel stability and the target vehicle health are input into the trained credit assessment model to obtain the credit score of the user to be assessed, including: Encoding the target driving style through an encoder of the trained credit assessment model to obtain a target driving style feature; The target travel stability is encoded by an encoder of the trained credit assessment model to obtain a target travel stability feature; The target vehicle health level is encoded by an encoder of the trained credit assessment model to obtain a target vehicle health level feature; The target driving style feature, the target travel stability feature and the target vehicle health feature are fused through a feature fusion layer of the trained credit assessment model to obtain a fusion feature; A target credit score matching the fusion feature is screened out from a plurality of preset credit scores, and the target credit score is determined as the credit score of the user to be evaluated.

3. The credit assessment method according to claim 2, characterized in that: The feature fusion layer of the trained credit assessment model fuses the target driving style feature, the target travel stability feature, and the target vehicle health feature to obtain a fusion feature, including: The feature fusion layer of the trained credit assessment model splices the target driving style features, the target travel stability features and the target vehicle health features in a preset splicing order to obtain a fusion feature.

4. The credit assessment method according to claim 2, characterized in that: The step of selecting a target credit score matching the fusion feature from a plurality of preset credit scores, and determining the target credit score as the credit score of the user to be evaluated includes: Determining the distribution probability of the fusion feature at each preset credit score; The preset credit score with the highest probability among the plurality of preset credit scores is determined as the credit score of the user to be evaluated.

5. The credit assessment method according to any one of claims 1 to 4, characterized in that: The step of inputting the historical driving data into the trained driving style prediction model to obtain a predicted target driving style of the user to be evaluated includes: encoding the historical driving data through an encoder of the trained driving style prediction model to obtain historical driving features; Determine the distribution probability of the historical driving characteristics in each preset driving style; The preset driving style with the largest distribution probability is determined as the target driving style of the user to be evaluated.

6. The credit assessment method according to any one of claims 1 to 4, characterized in that: The step of inputting the historical location data into the trained travel stability prediction model to obtain the predicted target travel stability of the user to be evaluated includes: The historical location data is encoded by an encoder of the trained travel stability prediction model to obtain historical travel characteristics; Determine the distribution probability of the historical travel characteristics at each preset travel stability level; The preset travel stability level with the largest distribution probability is determined as the target travel stability level of the user to be evaluated.

7. The credit assessment method according to any one of claims 1 to 4, characterized in that: The step of inputting the vehicle status data into the trained vehicle health assessment model to obtain an assessed target vehicle health level of the vehicle includes: The vehicle status data is encoded by an encoder of the trained vehicle health assessment model to obtain a vehicle status feature; Determining the distribution probability of the vehicle state characteristic at each preset vehicle health level; The preset vehicle health level with the largest distribution probability is determined as the target vehicle health level of the user to be evaluated.

8. A credit assessment device, characterized in that: include: An acquisition unit, used to acquire historical driving data, historical location data and vehicle status data of a vehicle corresponding to the user to be evaluated; A first input unit, configured to input the historical driving data into a trained driving style prediction model to obtain a predicted target driving style of the user to be evaluated; A second input unit is used to input the historical location data into the trained travel stability prediction model to obtain the predicted target travel stability of the user to be evaluated; A third input unit is used to input the vehicle status data into the trained vehicle health assessment model to obtain an assessed target vehicle health level of the vehicle; The fourth input unit is used to input the target driving style, the target travel stability and the target vehicle health into the trained credit assessment model to obtain the credit score of the user to be assessed.

9. An electronic device, characterized in that: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the credit assessment method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the credit assessment method according to any one of claims 1 to 7 is implemented.

Images