Vehicle owner character feature discrimination method and device, vehicle and storage medium

Abstract

The invention relates to a character feature discrimination method and device of a vehicle owner, a vehicle and a storage medium, and the method comprises the steps: obtaining the driving behavior data, portrait data and private domain community data of the vehicle owner; constructing a hybrid expert model according to a preset expert sub-model and a preset gating network model; and training a hybrid expert model by using the driving behavior data, the portrait data and the private domain community data to generate a trained hybrid expert model, analyzing the comprehensive feature data of the vehicle owner by using the trained hybrid expert model to generate to-be-discriminated feature data, and discriminating the character features of the vehicle owner according to the to-be-discriminated feature data. Therefore, the problems that in the prior art, fitting of all data generally depends on a single model, complex data sets with high heterogeneity are difficult to process, the generalization ability of the single model is limited, overfitting is prone to occurring, and efficiency is low when large-scale data are processed are solved.

Description

Technical Field

[0001] The present application relates to the technical field of personality analysis, and in particular to a method, device, vehicle and storage medium for determining the personality characteristics of a vehicle owner. Background Art

[0002] With the development of intelligent transportation systems and autonomous driving technology, the understanding and prediction of car owners' behavior has become increasingly important. The development of intelligent transportation systems and autonomous driving technology is an important trend in the transportation field. With the development of technologies such as cloud computing, big data, artificial intelligence, and the Internet of Things, they have become an important force in promoting economic growth and social development. In the development of intelligent transportation systems and autonomous driving technology, the analysis of car owners' personality characteristics has become increasingly important. This is because it not only involves the automated control of vehicles, but also involves the interaction with human driving behavior, as well as the understanding of car owners' behavior and preferences. For example, car owners with different personalities may have different acceptance, usage habits, and dependence on autonomous driving technology. Therefore, by analyzing the personality characteristics of car owners, we can better design and optimize the driving system and improve the user experience, which also helps to improve road traffic safety.

[0003] In related technologies, it is possible to fit data with the help of traditional machine learning models and generate a classification of users' actual population personality characteristics based on the driving behavior habit indicators of car owners.

[0004] However, related technologies usually rely on a single model to fit all data, which makes it difficult to handle complex and highly heterogeneous data sets. In addition, the generalization ability of a single model is limited, and it is prone to overfitting. It is inefficient when processing large-scale data and urgently needs to be improved. Summary of the invention

[0005] The present application provides a method, device, vehicle and storage medium for identifying the personality characteristics of a car owner, in order to solve the problems that the related technology usually relies on a single model to fit all data, is difficult to process complex and highly heterogeneous data sets, and the generalization ability of the single model is limited, it is easy to overfit, and has low efficiency when processing large-scale data.

[0006] The first aspect of the present application provides a method for identifying the personality characteristics of a car owner, comprising the following steps: acquiring the car owner's driving behavior data, portrait data, and private community data; constructing a hybrid expert model based on a preset expert sub-model and a preset gated network model; training the hybrid expert model using the driving behavior data, the portrait data, and the private community data to generate a trained hybrid expert model, and using the trained hybrid expert model to analyze the comprehensive feature data of the car owner to generate feature data to be identified, and identifying the personality characteristics of the car owner based on the feature data to be identified.

[0007] Optionally, in one embodiment of the present application, constructing a hybrid expert model based on a preset expert sub-model and a preset gating network model includes: establishing at least one expert sub-model among a long short-term memory network model, a clustering-logistic regression model and a large language model; and constructing the hybrid expert model based on the at least one expert sub-model and the preset gating network model.

[0008] Optionally, in one embodiment of the present application, before using the trained hybrid expert model to analyze the comprehensive feature data of the car owner, it also includes: inputting the driving behavior data, the portrait data and the private domain community data into the long short-term memory network model, the clustering-logistic regression model and the large language model respectively to generate the driving behavior feature representation, portrait feature representation and text language sentiment score of the car owner; weighted summing the driving behavior feature representation, the portrait feature representation, the text language sentiment score and the expert weights output by the gated network model to generate the comprehensive feature data of the car owner.

[0009] Optionally, in one embodiment of the present application, the use of the driving behavior data, the portrait data and the private domain community data to train the hybrid expert model to generate a trained hybrid expert model includes: using mean square error as the loss function, presetting the Adam algorithm as the optimizer and the mean absolute error as the evaluation index, dividing the driving behavior data, the portrait data and the private domain community data into a training set and a test set according to a preset ratio; training the hybrid expert model on the training set, and verifying the hybrid expert model on the test set until a preset stop condition is met, thereby generating the trained hybrid expert model.

[0010] Optionally, in one embodiment of the present application, the acquisition of the car owner's driving behavior data, portrait data and private community data includes: acquiring at least one of the vehicle identification code, speed changes, acceleration, braking frequency, steering habits, owner status, environmental perception data, cameras, lidar, weather changes and traffic density of the vehicle to generate the car owner's driving behavior data; acquiring at least one of the car owner's vehicle use habits, maintenance habits, consumption habits, car application usage habits and vehicle travel trajectory to generate the car owner's portrait data; acquiring at least one of the car owner's chats, comments and feedback to generate the car owner's private community data.

[0011] The second aspect of the present application provides a device for discriminating the personality characteristics of a car owner, including: an acquisition module for acquiring the car owner's driving behavior data, portrait data and private community data; a construction module for constructing a hybrid expert model based on a preset expert sub-model and a preset gated network model; a discrimination module for training the hybrid expert model using the driving behavior data, the portrait data and the private community data to generate a trained hybrid expert model, and using the trained hybrid expert model to analyze the comprehensive feature data of the car owner to generate feature data to be discriminated, and discriminate the personality characteristics of the car owner based on the feature data to be discriminated.

[0012] Optionally, in one embodiment of the present application, the construction module includes: an establishment unit for establishing at least one expert sub-model among the long short-term memory network model, the clustering-logistic regression model and the large language model; a construction unit for constructing the hybrid expert model based on the at least one expert sub-model and the preset gated network model.

[0013] Optionally, in one embodiment of the present application, it also includes: an input module, which is used to input the driving behavior data, the portrait data and the private domain community data into the long short-term memory network model, the clustering-logistic regression model and the large language model respectively before using the trained hybrid expert model to analyze the comprehensive feature data of the car owner, so as to generate the driving behavior feature representation, portrait feature representation and text language sentiment score of the car owner; a generation module, which is used to weighted sum the driving behavior feature representation, the portrait feature representation, the text language sentiment score and the expert weights output by the gated network model to generate the comprehensive feature data of the car owner.

[0014] Optionally, in one embodiment of the present application, the discrimination module includes: a division unit, used to use mean square error as the loss function, preset Adam algorithm as the optimizer and mean absolute error as the evaluation index, and divide the driving behavior data, the portrait data and the private domain community data into a training set and a test set according to a preset ratio; a training unit, used to train the hybrid expert model on the training set and verify the hybrid expert model on the test set until the preset stop condition is met, thereby generating the trained hybrid expert model.

[0015] Optionally, in one embodiment of the present application, the acquisition module includes: a first acquisition unit, used to acquire at least one of the vehicle identification code, speed changes, acceleration, braking frequency, steering habits, owner status, environmental perception data, cameras, lidar, weather changes and traffic density of the vehicle, and generate driving behavior data of the owner; a second acquisition unit, used to acquire at least one of the owner's vehicle use habits, maintenance habits, consumption habits, car application usage habits and vehicle travel trajectory, and generate portrait data of the owner; a third acquisition unit, used to acquire at least one of the owner's chats, comments and feedback, and generate private community data of the owner.

[0016] A third aspect of the present application provides a vehicle, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for identifying the personality characteristics of the vehicle owner as described in the above embodiment.

[0017] The fourth aspect of the present application provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the above method for determining the personality characteristics of the car owner.

[0018] The embodiment of the present application can collect the driving data and environmental factor data of the car owner, and construct, train and verify the hybrid expert model to identify the personality characteristics of the car owner, thereby reducing the risk of overfitting of a single model, improving the generalization ability and robustness of the overall model, and the hybrid expert model can be trained and predicted in parallel, improving computing efficiency, especially when processing large-scale data and complex tasks. Thus, it solves the problems that related technologies usually rely on a single model to fit all data, which is difficult to process complex and highly heterogeneous data sets, and the generalization ability of a single model is limited, it is easy to overfit, and the efficiency is low when processing large-scale data.

[0019] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0021] Figure 1 A flowchart of a method for identifying the personality characteristics of a car owner provided according to an embodiment of the present application;

[0022] Figure 2 is an overall flow chart of a method for identifying the personality characteristics of a car owner according to an embodiment of the present application;

[0023] Figure 3 The overall structure diagram of a method for identifying the personality characteristics of a car owner according to an embodiment of the present application;

[0024] Figure 4 This is a structural diagram of an LSTM model according to an embodiment of the present application;

[0025] Figure 5 A flow chart of a clustering model according to an embodiment of the present application;

[0026] Figure 6 A structure diagram of a logistic regression model according to an embodiment of the present application;

[0027] Figure 7 A Transformer model structure diagram according to an embodiment of the present application;

[0028] Figure 8 is a structural diagram of a gated network model according to an embodiment of the present application;

[0029] Fig. 9 A schematic diagram of the structure of a device for distinguishing the personality characteristics of a car owner provided according to an embodiment of the present application;

[0030] Fig.10 It is a schematic diagram of the structure of a vehicle provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0031] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0032] The following describes the method, device, vehicle and storage medium for distinguishing the personality characteristics of the car owner of the embodiment of the present application with reference to the accompanying drawings. In view of the problem that the related technologies mentioned in the above background technology usually rely on a single model to fit all data, it is difficult to process complex and highly heterogeneous data sets, and the generalization ability of the single model is limited, it is easy to overfit, and the efficiency is low when processing large-scale data, the present application provides a method for distinguishing the personality characteristics of the car owner, in which the driving data and environmental factor data of the car owner can be collected, and the hybrid expert model can be constructed, trained and verified to distinguish the personality characteristics of the car owner, thereby reducing the risk of overfitting of the single model, improving the generalization ability and robustness of the overall model, and the hybrid expert model can be trained and predicted in parallel, improving the computing efficiency, especially when processing large-scale data and complex tasks. Therefore, it solves the problems that the related technologies usually rely on a single model to fit all data, it is difficult to process complex and highly heterogeneous data sets, and the generalization ability of the single model is limited, it is easy to overfit, and the efficiency is low when processing large-scale data.

[0033] Specifically, Figure 1 A flowchart of a method for identifying the personality characteristics of a car owner provided in an embodiment of the present application.

[0034] like Figure 1 As shown, the method for determining the personality characteristics of the car owner includes the following steps:

[0035] In step S101, the car owner's driving behavior data, portrait data and private community data are obtained.

[0036] It can be understood that in the embodiment of the present application, the driving behavior data of the car owner may be the behavioral characteristics of the car owner during driving, the portrait data may be data generated based on the car owner's personal information and activity records, and the private community data may be data generated within the car owner's communication circle.

[0037] In the actual implementation process, Figure 2 As shown, the embodiments of the present application can obtain the driving behavior data, portrait data and private community data of the car owner, and pre-process the data, including data cleaning, feature extraction and normalization, so as to better capture the diversity and complexity of the data, improve the overall expression ability of the model, provide support for the training of subsequent models, and thus reduce occasional deviations.

[0038] Optionally, in one embodiment of the present application, the car owner's driving behavior data, portrait data and private community data are obtained, including: obtaining the vehicle identification code, speed changes, acceleration, braking frequency, steering habits, owner status, environmental perception data, cameras, lidar, weather changes and traffic density of the vehicle to generate the car owner's driving behavior data; obtaining at least one of the car owner's car use habits, maintenance habits, consumption habits, car application usage habits and vehicle travel trajectory to generate the car owner's portrait data; obtaining at least one of the car owner's chat, comments and feedback to generate the car owner's private community data.

[0039] It can be understood that in the embodiment of the present application, the driving behavior data of the car owner mainly comes from the TBOX (Telematics Box) and ADAS (Advanced Driving Assistance System) systems, the car owner's portrait data can be collected through CDP, and the car owner's private community data can be collected through SCRM (Social Customer Relationship Management).

[0040] In the actual implementation process, Figure 3 As shown, the embodiment of the present application can obtain driving behavior data such as VIN code, speed change, acceleration, braking frequency, steering habits, driver status monitoring, environmental perception, camera, lidar and possible environmental factors such as weather conditions, traffic density, etc. through TBOX and ADAS systems. The embodiment of the present application can obtain user portrait data such as car use habits, maintenance habits, consumption habits, travel trajectories, and car application usage habits through the CDP system. The embodiment of the present application can obtain user private domain community data such as car owners' chats, comments, feedback, etc. through the SCRM system.

[0041] The embodiments of the present application obtain the driving behavior data, portrait data and private community data of the car owner according to one or more methods, so as to further provide support for the training of subsequent models, thereby breaking the traditional questionnaire survey method to study the characteristics of the car owner and reducing occasional deviations.

[0042] It should be noted that data usage requires database support (Hive, SQLserver, etc.), and data processing analysis and algorithm application require support from pandas, numpy, sklearn, langchain, llamaindex, etc. in Python.

[0043] In step S102, a hybrid expert model is constructed according to a preset expert sub-model and a preset gating network model.

[0044] It is understandable that the Mixture of Experts (MoE) in the embodiment of the present application is an advanced neural network architecture that processes data by integrating multiple expert models. Each expert model can be a simple multi-layer perceptron (MLP) or a more complex model, such as a large language model (LLM). The MoE model includes multiple expert sub-models and a gated network model, and 3 expert sub-models are used in this application. As an advanced machine learning technology, the mixed expert model can process high-dimensional data and capture complex patterns in the data, providing new possibilities for the identification of the personality characteristics of car owners.

[0045] In the actual implementation process, the embodiment of the present application can construct a hybrid expert model based on the preset expert sub-model and the preset gating network model, thereby adopting a combination of multiple expert models, which can reduce the risk of overfitting of a single model, which not only improves the efficiency of the model, but also helps to improve the model's ability to handle complex tasks.

[0046] Optionally, in one embodiment of the present application, a hybrid expert model is constructed based on a preset expert sub-model and a preset gating network model, including: establishing at least one expert sub-model among a long short-term memory network model, a clustering-logistic regression model and a large language model; and constructing a hybrid expert model based on at least one expert sub-model and a preset gating network model.

[0047] It can be understood that the long short-term memory network LSTM model in the embodiment of the present application is a special RNN that can learn long-term dependencies, capture the dynamic changes of time series data, and extract the timing rules of input data. The number of neurons in the LSTM layer is 32, and the activation function is the Relu function. The large language LLM model in the embodiment of the present application is based on the Transformer architecture and outputs the user's text language according to the sentiment score.

[0048] Among them, the embodiment of the present application can construct a hybrid expert model based on expert sub-models such as long short-term memory network models, clustering-logistic regression models, and large language models and a preset gated network model, thereby further reducing the risk of overfitting of a single model based on the combination of multiple expert models.

[0049] Optionally, in one embodiment of the present application, before using the trained hybrid expert model to analyze the comprehensive feature data of the car owner, it also includes: inputting the driving behavior data, portrait data and private domain community data into the long short-term memory network model, the clustering-logistic regression model and the large language model respectively to generate the car owner's driving behavior feature representation, portrait feature representation and text language sentiment score; weighted summing the driving behavior feature representation, portrait feature representation, text language sentiment score and the expert weights output by the gated network model to generate the car owner's comprehensive feature data.

[0050] It is understandable that if Figure 4 As shown, the core of the LSTM model in the embodiment of the present application lies in its unit structure. Each unit contains four main parts: (1) Input gate: determines which parts of the input information at the current moment are allowed to enter the state unit. (2) Forget gate: determines which information in the state unit needs to be forgotten or removed. (3) Cell state: a core component for storing historical state information. (4) Output gate: determines the information output to the next hidden layer. (5) Activation function (Sigmoid, Tanh): used for nonlinear transformation. (6) Output layer: LSTM output data corresponding to the input data. LSTM controls the flow of information entry, retention and output by opening and closing the input gate, forget gate and output gate, thereby realizing the memory and processing of time series information. Compared with simple RNN, LSTM can better capture long-term dependencies in time series data, process longer time series, have more powerful modeling capabilities, and generate driving behavior feature representations.

[0051] Furthermore, if Figure 5 and Figure 6 As shown, the user portrait adopts a clustering-logistic regression model, the clustering algorithm adopts the K-Means algorithm (K=5 can be set first), and the activation function of the logistic regression is the Sigmoid function, which is used to perform nonlinear transformation and generate a portrait feature representation.

[0052] The logistic regression model predicts the probability of an event by mapping the output of the linear regression model to between 0 and 1 using the Sigmoid function (also called the logistic function). The formula for the Sigmoid function is as follows:

[0053]

[0054] Where z is a linear combination, usually a weighted sum of the input features, and e is the base of the natural logarithm. The output of this function is between 0 and 1 and can be interpreted as a probability.

[0055] The basic form of the logistic regression model can be expressed as:

[0056]

[0057] Among them, P(Y=1|X) is the probability that the target variable Y is equal to 1 given the input feature X, and β0, β1,..., βn are model parameters.

[0058] Furthermore, if Figure 7As shown, the user private domain data analysis in the embodiment of the present application adopts the LLM model, based on the Transformer architecture, to output the user's text language according to the sentiment score.

[0059] Clean and format the collected chats, comments, and feedback. This usually includes removing irrelevant characters (such as HTML tags), special symbols, stop words, and performing word segmentation. The purpose of preprocessing is to extract features that are helpful for sentiment analysis. Convert the text into a numerical representation that the model can handle. Common methods include Bag-of-Words, TF-IDF (Term Frequency-Inverse Document Frequency), and Word Embedding. These methods can convert text into vector form and retain the semantic information of the vocabulary.

[0060] The embodiments of the present application can select a pre-trained large language model, such as BERT, GPT series, etc. These models have been pre-trained on a large amount of text data and can capture the complex patterns and semantic information of the language. Before using the pre-trained LLM, it is usually necessary to fine-tune it on a training data set labeled with sentiment categories. In this way, the model can learn features and patterns specific to sentiment analysis tasks. The processed chat, comments, and feedback content are input into the fine-tuned LLM, and the model will output the sentiment probability distribution of each category. According to the category with the highest output probability, the sentiment tendency and score of the chat content are determined.

[0061] Furthermore, if Figure 8 As shown in the figure, the gated network model adopts a multi-layer perceptron model. It is a 3-layer network structure, the input layer receives input data, the middle layer is a 32-dimensional vector, the activation function is the relu function, and the output layer is a 3-dimensional vector, the activation function is softmax, which corresponds to the weights of the 3 expert sub-models, and the sum of the 3 weights is 1.

[0062] The embodiment of the present application can perform weighted summation, and the outputs of the three expert sub-models are output1, output2 and output3 respectively. These three outputs are all high-dimensional vectors. Taking 32 dimensions as an example, they respectively represent the feature representation of each sub-expert model for the input data. The feature representation output by the three expert sub-models is weightedly summed with the expert weights output by the gating network to obtain comprehensive features, which can be trained and predicted in parallel to improve computing efficiency, especially when processing large-scale data and complex tasks.

[0063] It should be noted that the use of LLM large language models includes llama model, ChatGPT, chatGLM, Qwen, kimi, etc.

[0064] In step S103, the hybrid expert model is trained using the driving behavior data, the portrait data, and the private domain community data to generate a trained hybrid expert model, and the trained hybrid expert model is used to analyze the comprehensive feature data of the car owner to generate feature data to be judged, and the personality characteristics of the car owner are judged based on the feature data to be judged.

[0065] It can be understood that the personality traits in the embodiments of the present application include but are not limited to responsible personality traits (such as the owner maintains the vehicle and obeys traffic rules), extroverted personality traits (such as the owner is active on social media) and unstable personality traits (such as the owner has strong emotional reactions when there is traffic congestion).

[0066] In the actual implementation process, the embodiment of the present application can use driving behavior data, portrait data and private community data to train a hybrid expert model to generate a trained hybrid expert model. After the hybrid expert model is trained and the model effect is good, the trained hybrid expert model is used to analyze the comprehensive feature data of the car owner to generate feature data to be judged, deploy the hybrid expert model, and use the hybrid expert model to predict the car owner's personality score.

[0067] The embodiment of the present application can use the trained hybrid expert model to analyze the comprehensive feature data of the car owner to determine the personality characteristics of the car owner, thereby improving the generalization ability and robustness of the overall model, allowing the large model to understand and select the correct application to execute.

[0068] Optionally, in one embodiment of the present application, a hybrid expert model is trained using driving behavior data, portrait data, and private domain community data to generate a trained hybrid expert model, including: using mean square error as the loss function, presetting the Adam algorithm as the optimizer and the mean absolute error as the evaluation indicator, dividing the driving behavior data, portrait data, and private domain community data into a training set and a test set according to a preset ratio; training the hybrid expert model on the training set, and verifying the hybrid expert model on the test set until the preset stop condition is met, thereby generating a trained hybrid expert model.

[0069] It is understandable that the preset ratio in the embodiment of the present application may be 8: 2. The activation function in the embodiment of the present application is a linear function, and the final output value is the owner's personality score.

[0070] In the actual execution process, the embodiment of the present application implements the MoE model code and trains based on the TensorFlow framework. The number of training epochs = 100 and the number of batch samples batch_size = 32. The optimizer for model training is the adam algorithm, the loss function is the mean square error (mse), and the evaluation indicator metrics is the mean absolute error (MAE), that is, the sum of the absolute values ​​of the differences between the target value and the predicted value. The driving behavior data, portrait data, and private domain community data are divided into training sets and test sets in a ratio of 8:2. The hybrid expert model is trained on the training set and verified on the test set until the stopping condition is met, the training is stopped, and the trained hybrid expert model is generated.

[0071] The embodiment of the present application can provide support for analyzing the comprehensive feature data of the car owner and judging the car owner's personality based on the trained hybrid expert model, thereby providing personalized services based on the car owner's personality characteristics, ensuring safe driving assistance for the vehicle, and improving the user experience.

[0072] It should be noted that the preset stop condition can be set by those skilled in the art according to actual conditions and is not specifically limited here.

[0073] According to the method for distinguishing the personality characteristics of the car owner proposed in the embodiment of the present application, the driving data and environmental factor data of the car owner can be collected to construct, train and verify the hybrid expert model, so as to distinguish the personality characteristics of the car owner, thereby reducing the risk of overfitting of a single model, improving the generalization ability and robustness of the overall model, and the hybrid expert model can be trained and predicted in parallel, improving computing efficiency, especially when processing large-scale data and complex tasks. Thus, the problem that the relevant technology usually relies on a single model to fit all data, is difficult to process complex and highly heterogeneous data sets, and the generalization ability of a single model is limited, it is easy to overfit, and has low efficiency when processing large-scale data is solved.

[0074] Next, the device for distinguishing the personality characteristics of a car owner proposed in accordance with an embodiment of the present application will be described with reference to the accompanying drawings.

[0075] Fig. 9 It is a schematic diagram of the structure of the device for distinguishing the personality characteristics of the car owner according to an embodiment of the present application.

[0076] like Fig. 9 As shown, the device 10 for determining the personality characteristics of the vehicle owner includes: an acquisition module 100 , a construction module 200 and a determination module 300 .

[0077] Specifically, the acquisition module 100 is used to obtain the car owner's driving behavior data, portrait data and private community data.

[0078] The construction module 200 is used to construct a hybrid expert model according to a preset expert sub-model and a preset gating network model.

[0079] The discrimination module 300 is used to train the hybrid expert model using driving behavior data, portrait data and private community data to generate a trained hybrid expert model, and use the trained hybrid expert model to analyze the comprehensive feature data of the car owner to generate feature data to be discriminated, and discriminate the personality characteristics of the car owner based on the feature data to be discriminated.

[0080] Optionally, in one embodiment of the present application, the construction module 200 includes: an establishment unit and a construction unit.

[0081] Among them, the establishment unit is used to establish at least one expert sub-model among the long short-term memory network model, the clustering-logistic regression model and the large language model.

[0082] A construction unit is used to construct a hybrid expert model according to at least one expert sub-model and a preset gating network model.

[0083] Optionally, in one embodiment of the present application, the device 10 for determining the personality characteristics of the vehicle owner further includes: an input module and a generation module.

[0084] Among them, the input module is used to input the driving behavior data, portrait data and private community data into the long short-term memory network model, the clustering-logistic regression model and the large language model respectively before using the trained hybrid expert model to analyze the comprehensive feature data of the car owner, so as to generate the car owner's driving behavior feature representation, portrait feature representation and text language sentiment score.

[0085] The generation module is used to weightedly sum the driving behavior feature representation, portrait feature representation, text language sentiment score and expert weights output by the gated network model to generate comprehensive feature data of the car owner.

[0086] Optionally, in one embodiment of the present application, the discrimination module 300 includes: a division unit and a training unit.

[0087] Among them, the division unit is used to use the mean square error as the loss function, preset the Adam algorithm as the optimizer and the mean absolute error as the evaluation indicator, and divide the driving behavior data, portrait data and private community data into training sets and test sets according to the preset ratio.

[0088] The training unit is used to train the hybrid expert model on the training set and verify the hybrid expert model on the test set until a preset stop condition is met, thereby generating a trained hybrid expert model.

[0089] Optionally, in one embodiment of the present application, the acquisition module 100 includes: a first acquisition unit, a second acquisition unit and a third acquisition unit.

[0090] Among them, the first acquisition unit is used to obtain at least one of the vehicle identification code, speed change, acceleration, braking frequency, steering habits, owner status, environmental perception data, camera, lidar, weather changes and traffic density of the vehicle to generate the owner's driving behavior data.

[0091] The second acquisition unit is used to obtain at least one of the car owner's car usage habits, maintenance habits, consumption habits, car application usage habits and vehicle travel trajectory, and generate portrait data of the car owner.

[0092] The third acquisition unit is used to obtain at least one of the car owner's chat, comments and feedback to generate the car owner's private community data.

[0093] It should be noted that the above explanation of the embodiment of the method for determining the personality characteristics of a vehicle owner is also applicable to the device for determining the personality characteristics of a vehicle owner of this embodiment, and will not be repeated here.

[0094] According to the device for distinguishing the personality characteristics of the car owner proposed in the embodiment of the present application, the driving data and environmental factor data of the car owner can be collected to construct, train and verify the hybrid expert model, so as to distinguish the personality characteristics of the car owner, thereby reducing the risk of overfitting of a single model, improving the generalization ability and robustness of the overall model, and the hybrid expert model can be trained and predicted in parallel, improving computing efficiency, especially when processing large-scale data and complex tasks. Thus, the problem that the related technology usually relies on a single model to fit all data, is difficult to process complex and highly heterogeneous data sets, and the generalization ability of a single model is limited, it is easy to overfit, and is inefficient when processing large-scale data is solved.

[0095] Fig.10 A schematic diagram of the structure of a vehicle provided in an embodiment of the present application. The vehicle may include:

[0096] A memory 1001 , a processor 1002 , and a computer program stored in the memory 1001 and executable on the processor 1002 .

[0097] When the processor 1002 executes the program, the method for determining the personality characteristics of the vehicle owner provided in the above embodiment is implemented.

[0098] Furthermore, the vehicle also includes:

[0099] The communication interface 1003 is used for communication between the memory 1001 and the processor 1002 .

[0100] The memory 1001 is used to store computer programs that can be executed on the processor 1002 .

[0101] The memory 1001 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0102] If the memory 1001, the processor 1002 and the communication interface 1003 are implemented independently, the communication interface 1003, the memory 1001 and the processor 1002 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig.10 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0103] Optionally, in a specific implementation, if the memory 1001, the processor 1002 and the communication interface 1003 are integrated on a chip, the memory 1001, the processor 1002 and the communication interface 1003 can communicate with each other through an internal interface.

[0104] The processor 1002 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0105] This embodiment also provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the above method for determining the personality characteristics of a vehicle owner is implemented.

[0106] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0107] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0108] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0109] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0110] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0111] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0112] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0113] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for identifying the personality characteristics of a car owner, characterized in that: The following steps are involved: Obtain the car owner's driving behavior data, portrait data and private community data; Constructing a hybrid expert model according to a preset expert sub-model and a preset gating network model; The hybrid expert model is trained using the driving behavior data, the portrait data, and the private domain community data to generate a trained hybrid expert model, and the trained hybrid expert model is used to analyze the comprehensive feature data of the car owner to generate feature data to be determined, and the personality characteristics of the car owner are determined based on the feature data to be determined.

2. The method according to claim 1, characterized in that The step of constructing a hybrid expert model according to a preset expert sub-model and a preset gating network model includes: Establish at least one expert sub-model among the long short-term memory network model, the clustering-logistic regression model and the large language model; The hybrid expert model is constructed according to the at least one expert sub-model and the preset gating network model.

3. The method according to claim 2, characterized in that Before using the trained hybrid expert model to analyze the comprehensive feature data of the vehicle owner, the method further includes: Inputting the driving behavior data, the portrait data and the private domain community data into the long short-term memory network model, the clustering-logistic regression model and the large language model respectively to generate the driving behavior feature representation, the portrait feature representation and the text language sentiment score of the car owner; The driving behavior feature representation, the portrait feature representation, the text language sentiment score and the expert weight output by the gated network model are weighted and summed to generate comprehensive feature data of the vehicle owner.

4. The method according to claim 1, characterized in that: The using the driving behavior data, the portrait data and the private domain community data to train the hybrid expert model to generate a trained hybrid expert model includes: Using mean square error as the loss function, Adam algorithm as the optimizer and mean absolute error as the evaluation index, the driving behavior data, the portrait data and the private domain community data are divided into a training set and a test set according to a preset ratio; The hybrid expert model is trained on the training set, and the hybrid expert model is verified on the test set until a preset stop condition is met, thereby generating the trained hybrid expert model.

5. The method according to claim 1, characterized in that The acquisition of the car owner's driving behavior data, portrait data and private community data includes: Obtaining at least one of a vehicle identification code, speed change, acceleration, braking frequency, steering habits, owner status, environmental perception data, camera, lidar, weather change, and traffic density of the vehicle to generate driving behavior data of the owner; Obtain at least one of the vehicle owner's vehicle use habits, maintenance habits, consumption habits, vehicle computer application usage habits, and vehicle travel trajectory, and generate portrait data of the vehicle owner; At least one of the chat, comment and feedback of the car owner is obtained to generate the private domain community data of the car owner.

6. A device for identifying the personality characteristics of a car owner, characterized in that: include: The acquisition module is used to obtain the car owner's driving behavior data, portrait data and private community data; A construction module, used for constructing a hybrid expert model according to a preset expert sub-model and a preset gating network model; A discrimination module is used to train the hybrid expert model using the driving behavior data, the portrait data and the private domain community data to generate a trained hybrid expert model, and use the trained hybrid expert model to analyze the comprehensive feature data of the car owner to generate feature data to be discriminated, and discriminate the personality characteristics of the car owner based on the feature data to be discriminated.

7. The device according to claim 6, characterized in that The building blocks include: A building unit, used to build at least one expert sub-model among a long short-term memory network model, a clustering-logistic regression model and a large language model; A construction unit is used to construct the hybrid expert model according to the at least one expert sub-model and the preset gating network model.

8. The device according to claim 7, characterized in that Also includes: An input module, used for inputting the driving behavior data, the portrait data and the private domain community data into the long short-term memory network model, the clustering-logistic regression model and the large language model respectively before analyzing the comprehensive feature data of the car owner using the trained hybrid expert model, so as to generate the driving behavior feature representation, the portrait feature representation and the text language sentiment score of the car owner; A generation module is used to weightedly sum the driving behavior feature representation, the portrait feature representation, the text language sentiment score and the expert weight output by the gated network model to generate comprehensive feature data of the car owner.

9. A vehicle, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for identifying the personality characteristics of a vehicle owner as described in any one of claims 1 to 5.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the method for identifying the personality characteristics of a vehicle owner as described in any one of claims 1 to 5.

Images