Charge evasion behavior inspection method and device, computer equipment and readable storage medium

Abstract

The invention relates to a fee evasion behavior inspection method and device, computer equipment and a readable storage medium. The method comprises the steps of obtaining a plurality of preset fee evasion features related to passing of an ETC lane, and determining an initial feature vector of a target vehicle for each preset fee evasion feature; acquiring respective preset fee evasion weights of the plurality of preset fee evasion features; determining a target feature vector of the target vehicle for each preset fee evasion feature according to a preset fee evasion weight and the initial feature vector of the target vehicle for each preset fee evasion feature; and based on the target feature vector of the target vehicle for each preset fee evasion feature, determining the probability that the target vehicle has a fee evasion behavior when passing through the ETC lane. With the adoption of the method, the speed and efficiency of identifying the toll evasion vehicle can be improved, the potential toll evasion behavior can be warned in advance, an advanced inspection means is provided, and the post-treatment cost and difficulty are reduced.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, apparatus, computer equipment, computer-readable storage medium and computer program product for investigating fee evasion. Background Art

[0002] The ETC system has the advantages of being fast, safe, environmentally friendly and efficient. In particular, the ETC network has been realized on national highways. More and more car owners choose to apply for and install ETC cards and use the ETC system to pass through highway toll stations, which greatly improves people's travel efficiency, especially during holidays and weekends.

[0003] However, although ETC non-stop toll collection has alleviated traffic problems to a certain extent, some customers who evade tolls have taken advantage of this convenience to evade tolls. The large amount of tolls lost has attracted the attention of highway management departments, and they have begun to conduct strict inspections on passing vehicles. The current inspection method is mostly based on manual on-site inspections, supplemented by background inspections. The former is not only inefficient, but also has little detection, while the latter is lagging and has not played a good deterrent role. Summary of the invention

[0004] Based on this, it is necessary to provide a method, device, computer equipment, computer-readable storage medium and computer program product for investigating toll evasion behavior that can improve the accuracy and timeliness of high-speed ETC inspections in response to the above-mentioned technical problems.

[0005] In a first aspect, the present application provides a method for investigating fee evasion, comprising:

[0006] Acquire multiple preset toll evasion features related to the ETC lane, and determine an initial feature vector of the target vehicle for each of the preset toll evasion features;

[0007] Obtaining a preset fee evasion weight for each of the plurality of preset fee evasion features;

[0008] Determining a target feature vector of the target vehicle for each of the preset fare evasion features according to the preset fare evasion weight and the initial feature vector of the target vehicle for each of the preset fare evasion features;

[0009] Based on the target feature vector of the target vehicle for each of the preset toll evasion features, the probability that the target vehicle has toll evasion behavior when passing through the ETC lane is determined.

[0010] In one embodiment, the determining the probability that the target vehicle has evaded tolls in the ETC lane based on the target feature vector of the target vehicle for each of the preset evasion features includes:

[0011] Obtain a preset slope parameter corresponding to each of the preset evasion features; for each of the preset evasion features, multiply the preset slope parameter of the preset evasion feature with the target feature vector to obtain the feature vector to be accumulated corresponding to the preset evasion feature; accumulate the feature vectors to be accumulated corresponding to each of the preset evasion features to obtain a cumulative feature vector; obtain a preset intercept parameter, and based on the preset intercept parameter and the cumulative feature vector, determine the probability that the target vehicle has evaded the fee in the ETC lane.

[0012] In one embodiment, the obtaining of a preset intercept parameter and determining the probability that the target vehicle has evaded tolls in the ETC lane based on the preset intercept parameter and the accumulated and characteristic vector include:

[0013] The preset intercept parameter and the cumulative sum feature vector are added to obtain a linear feature vector; the linear feature vector is exponentially transformed to obtain a nonlinear feature vector; based on the nonlinear feature vector, the probability that the target vehicle has evaded tolls in the ETC lane is determined.

[0014] In one of the embodiments, a preset probability threshold is obtained, and the probability that the target vehicle has evaded tolls in the ETC lane is compared with the preset probability threshold; when the probability that the target vehicle has evaded tolls in the ETC lane is greater than or equal to the preset probability threshold, the evasion behavior of the target vehicle in the ETC lane is classified based on the target feature vector of the target vehicle for each of the preset evasion features.

[0015] In one embodiment, the classifying the toll evasion behavior of the target vehicle in the ETC lane based on the target feature vector of the target vehicle for each of the preset toll evasion features includes:

[0016] Obtain a first weight value and a first bias value corresponding to the evasion behavior classification category; for each of the preset evasion features, based on the first weight value and the first bias value, convert the target feature vector for the preset evasion feature into an activation value corresponding to the evasion behavior classification category; based on the activation value corresponding to the evasion behavior classification category, obtain the evasion behavior probability distribution feature vector for the preset evasion feature; classify the evasion behavior of the target vehicle in the ETC lane according to the evasion behavior probability distribution feature vector corresponding to each of the preset evasion features.

[0017] In one embodiment, the obtaining of a plurality of preset toll evasion features related to passing through an ETC lane includes:

[0018] Acquire multiple candidate evasion features related to the ETC lane; for each of the candidate evasion features, output the candidate evasion feature to an evasion decision tree, the evasion decision tree including at least one evasion node; determine the impurity reduction amount caused by the splitting of the candidate evasion feature when passing through each evasion node in the evasion decision tree; based on the impurity reduction amount caused by the splitting when passing through each evasion node, determine the average impurity reduction amount of the candidate evasion feature; based on the average impurity reduction amount corresponding to each of the multiple candidate evasion features, screen out multiple preset evasion features from the multiple candidate evasion features.

[0019] In a second aspect, the present application also provides a device for checking fee evasion, including:

[0020] A first determination module is used to obtain a plurality of preset evasion features related to the ETC lane, and determine an initial feature vector of the target vehicle for each of the preset evasion features;

[0021] An acquisition module, configured to acquire a preset evasion weight of each of the plurality of preset evasion features;

[0022] A second determination module is used to determine a target feature vector of the target vehicle for each of the preset fare evasion features according to the preset fare evasion weight and the initial feature vector of the target vehicle for each of the preset fare evasion features;

[0023] The third determination module is used to determine the probability that the target vehicle has evaded tolls when passing through the ETC lane based on the target feature vector of the target vehicle for each of the preset evasion features.

[0024] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0025] Acquire multiple preset toll evasion features related to the ETC lane, and determine an initial feature vector of the target vehicle for each of the preset toll evasion features;

[0026] Obtaining a preset fee evasion weight for each of the plurality of preset fee evasion features;

[0027] Determining a target feature vector of the target vehicle for each of the preset fare evasion features according to the preset fare evasion weight and the initial feature vector of the target vehicle for each of the preset fare evasion features;

[0028] Based on the target feature vector of the target vehicle for each of the preset toll evasion features, the probability that the target vehicle has toll evasion behavior when passing through the ETC lane is determined.

[0029] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0030] Acquire multiple preset toll evasion features related to the ETC lane, and determine an initial feature vector of the target vehicle for each of the preset toll evasion features;

[0031] Obtaining a preset fee evasion weight for each of the plurality of preset fee evasion features;

[0032] Determining a target feature vector of the target vehicle for each of the preset fare evasion features according to the preset fare evasion weight and the initial feature vector of the target vehicle for each of the preset fare evasion features;

[0033] Based on the target feature vector of the target vehicle for each of the preset toll evasion features, the probability that the target vehicle has toll evasion behavior when passing through the ETC lane is determined.

[0034] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0035] Acquire multiple preset toll evasion features related to the ETC lane, and determine an initial feature vector of the target vehicle for each of the preset toll evasion features;

[0036] Obtaining a preset fee evasion weight for each of the plurality of preset fee evasion features;

[0037] Determining a target feature vector of the target vehicle for each of the preset fare evasion features according to the preset fare evasion weight and the initial feature vector of the target vehicle for each of the preset fare evasion features;

[0038] Based on the target feature vector of the target vehicle for each of the preset toll evasion features, the probability that the target vehicle has toll evasion behavior when passing through the ETC lane is determined.

[0039] The above-mentioned evasion behavior inspection method, device, computer equipment, computer-readable storage medium and computer program product can comprehensively capture various possible evasion behavior patterns by obtaining multiple preset evasion characteristics related to the ETC lanes and determining the initial feature vector of the target vehicle for each preset evasion characteristic. It provides basic data support for subsequent data analysis, making it easier to identify and distinguish between normal passing vehicles and potential evasion vehicles, thereby improving the pertinence and accuracy of the inspection.

[0040] Obtaining the preset evasion weights of the plurality of preset evasion features and assigning weights to different evasion features can highlight factors that have a greater impact on evasion behavior, ensuring that more attention is paid to important features during the evaluation process. This helps to improve the accuracy of the prediction model, reduce the false alarm rate, be closer to the actual situation, and enhance its applicability and reliability.

[0041] According to the preset evasion weight and the initial feature vector of the target vehicle for each preset evasion feature, the target feature vector of the target vehicle for each preset evasion feature is determined, and the weight is applied to the initial feature vector to generate the target feature vector. In fact, the input data is adjusted and optimized to better reflect the actual evasion risk. The ability to understand evasion behavior is enhanced, making the prediction result more accurate, and also increasing the adaptability to new situations.

[0042] Based on the target feature vector of the target vehicle for each of the preset evasion features, the probability of the target vehicle evading fees in the ETC lane is determined. The final step is to use the weighted optimized feature vector to calculate the evasion probability, which directly outputs the result of whether the vehicle is likely to evade fees. This method not only improves the speed and efficiency of identifying evasion vehicles, but also can warn of potential evasion behaviors in advance, provide a pre-emptive audit method, and reduce the cost and difficulty of post-processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 A diagram showing an application environment of a method for investigating fee evasion in an embodiment;

[0045] Figure 2 A schematic diagram of a flow chart of a method for investigating fee evasion in one embodiment;

[0046] Figure 3 A schematic diagram of a framework of a method for investigating fee evasion in one embodiment;

[0047] Figure 4 A schematic diagram of a flow chart of a method for investigating fee evasion in another embodiment;

[0048] Figure 5 It is a structural block diagram of a device for investigating fee evasion in one embodiment;

[0049] Figure 6 It is a structural block diagram of a device for investigating fee evasion in another embodiment;

[0050] Figure 7 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0051] 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.

[0052] The fee evasion behavior inspection method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The terminal 102 is used to generate a fee evasion behavior audit request and send the fee evasion behavior audit request to the server 104, so that the server 104 determines the probability that the target vehicle has a fee evasion behavior in the ETC lane based on the target feature vector of the target vehicle for each preset fee evasion feature. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car devices, projection devices, etc. The portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, etc. The head-mounted device can be a virtual reality (VR) device, an augmented reality (AR) device, smart glasses, etc. The server 104 may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services.

[0053] In an exemplary embodiment, Figure 2 As shown, a method for checking fee evasion behavior is provided, and the method is applied to Figure 1 The server 104 in the example is used as an example to illustrate, including the following steps 202 to 208. Among them:

[0054] Step 202, obtaining a plurality of preset toll evasion features related to the ETC lane, and determining an initial feature vector of the target vehicle for each preset toll evasion feature.

[0055] Among them, ETC lanes refer to dedicated lanes for electronic toll collection systems, which allow vehicles equipped with on-board units (OBU) and ETC cards to automatically pay tolls without stopping when passing through toll booths. Understandably, ETC lanes are usually equipped with special reading and writing devices that can identify the OBU on the vehicle and communicate with it to achieve a fast, safe and efficient toll collection process.

[0056] Preset evasion features refer to attributes or indicators that can indicate that a vehicle may be evading fees. These features may include vehicle model mismatch, vehicle card mismatch, mobile OBU equipment, truck OBU missed, opposite-direction signature replacement, malicious vehicle following, false card collection, no card at the exit, U-turn evasion, etc.

[0057] The target vehicle is a specific vehicle that is being evaluated by the audit system for toll evasion. Understandably, the target vehicle will pass through the ETC lane, and its travel information will be collected and used as a data source for analysis.

[0058] The initial feature vector refers to the conversion of the associated attributes or states of each individual preset toll evasion feature into a set of quantitative values ​​that are easy to process and analyze based on the specific situation of the target vehicle.

[0059] Specifically, firstly, based on historical data analysis and known evasion behavior patterns, a series of features that can characterize possible evasion behaviors are determined. These features can be vehicle type mismatch, vehicle card mismatch, mobile OBU devices, etc. Ensure that all potential evasion behavior types can be captured, thereby fully covering the audit needs.

[0060] Then, every time a vehicle (target vehicle) passes through the ETC lane, relevant information of the target vehicle is automatically recorded, such as vehicle model, number of axles, entry and exit time, total weight, etc. This information will serve as the basic data for subsequent analysis, providing the necessary raw data support for each preset toll evasion feature for quantitative processing.

[0061] Finally, for each preset evasion feature selected, it is converted into one or more quantifiable indicators. For example, the vehicle model code is converted into a virtual variable, or a continuous value such as time interval or weight measured by a sensor is directly used. Finally, for each preset evasion feature, an initial feature vector is constructed according to its corresponding quantification result. This vector contains all relevant quantitative information of the target vehicle on this feature.

[0062] In one example, for a preset toll evasion feature such as "model mismatch", the initial feature vector may be [0, 1], indicating that the actual model of the target vehicle is a small passenger car, but the OBU is set to a truck.

[0063] Step 204: Obtain preset evasion weights of respective preset evasion features.

[0064] The preset evasion weight refers to a numerical value assigned to each preset evasion feature, which is used to indicate the importance or influence of the feature in determining whether the vehicle has evaded fare. The purpose is to reflect the different contributions of different features to the identification of evasion behavior.

[0065] Specifically, first collect a large amount of historical traffic records, i.e., raw data, including known cases of toll evasion and normal traffic records. Further clean the raw data, process missing values, error values, and delete useless variables. Extract valid feature attribute information fields, such as vehicle model, number of axles, import and export time, total weight, etc. Virtualize type variables (such as vehicle model) and normalize continuous value variables (such as total weight). Understandably, the purpose of this step is to ensure the quality and consistency of data used for weight calculation and improve the effectiveness of model training.

[0066] Then, the selected weight training model (such as the random forest model) is applied to train the processed data. The feature importance scores output by the weight training model are analyzed as the basis for preliminary weight allocation. The parameters are adjusted through multiple iterations to optimize the model performance until a satisfactory feature weight is obtained. The feature weight is used to provide a stable and verified weight configuration for future evasion behavior prediction.

[0067] Step 206 , determining a target feature vector of the target vehicle for each preset fare evasion feature according to the preset fare evasion weight and the initial feature vector of the target vehicle for each preset fare evasion feature.

[0068] The target feature vector is a quantitative representation of the state of the target vehicle for each preset evasion feature after combining the preset evasion weights. Specifically, it is a new feature vector formed by introducing weights to adjust the importance of each feature on the basis of the initial feature vector. This process can be regarded as a weighted processing of the original data, with the purpose of making those features that are more important for identifying evasion behavior occupy a larger proportion in the final analysis.

[0069] Specifically, each preset evasion feature and its corresponding initial feature vector are traversed. For each preset evasion feature, the following steps are performed:

[0070] The preset evasion weight corresponding to the preset evasion feature is taken out, and each element in the initial feature vector is multiplied by the corresponding weight value to obtain the weighted feature value, and the weighted feature value is stored in the previously initialized target feature vector container. Through the above feature-by-feature processing, it is ensured that each preset evasion feature is correctly assigned a weight and updated to the target feature vector.

[0071] In one example, assuming that the preset evasion feature is the vehicle type, and the preset evasion weight corresponding to the preset evasion feature is 0.3. Then for a specific target vehicle, the data for the preset evasion feature is [small passenger car], and the preset evasion feature is further converted into a quantifiable indicator. Assuming that the quantifiable indicator is [1,0] (binary representation), after combining the weight, the target feature vector obtained is [1×0.3, 1×0.3].

[0072] Step 208 , based on the target feature vector of the target vehicle for each preset toll evasion feature, determine the probability that the target vehicle has toll evasion behavior in the ETC lane.

[0073] Toll evasion refers to the behavior of vehicles taking improper means to avoid paying tolls when passing through the ETC (Electronic Toll Collection System) lanes. Such behavior violates normal toll collection rules, resulting in the highway operator failing to receive the tolls it deserves, thus causing economic losses and management problems.

[0074] Specifically, a toll evasion prediction model (such as a certain logistic regression model) is used to predict whether the target vehicle has engaged in toll evasion in the ETC lane. Specifically, for each preset toll evasion feature of the target vehicle, the target feature vector corresponding to each preset toll evasion feature is input into the toll evasion prediction model, and the target feature vector corresponding to each preset toll evasion feature is processed by the toll evasion prediction model to obtain a probability result output by the toll evasion prediction model, which reflects the probability that the target vehicle has engaged in toll evasion.

[0075] In one example, the probability of evasion behavior is predicted by the evasion prediction model. For the preset evasion feature of the target vehicle, i.e., the vehicle model, the target feature vector corresponding to the preset evasion feature may be [0, 1]. After inputting this target feature vector into the evasion prediction model, the evasion prediction model outputs a result of 0.75, which means that there is a 75% probability that the target vehicle will evade the fee.

[0076] In one of the embodiments, a preset slope parameter corresponding to each preset evasion feature is obtained; for each preset evasion feature, the preset slope parameter of the preset evasion feature and the target feature vector are multiplied to obtain the feature vector to be accumulated corresponding to the preset evasion feature; the feature vector to be accumulated corresponding to each preset evasion feature is accumulated to obtain a cumulative feature vector; a preset intercept parameter is obtained, and based on the preset intercept parameter and the cumulative feature vector, the probability of the target vehicle engaging in evasion behavior in the ETC lane is determined.

[0077] The preset slope parameter refers to the weight or coefficient in the logistic regression model. Each preset evasion feature has a corresponding slope parameter, which determines the degree of influence of the feature on the final prediction result.

[0078] The feature vector to be accumulated refers to the result obtained by multiplying a preset evasion feature with its corresponding preset slope parameter. It can be understood that for each preset evasion feature, there will be such a vector, which represents the contribution of the preset evasion feature to the evasion probability prediction.

[0079] The accumulated feature vector refers to the sum of the results of multiplying each preset evasion feature with their respective preset slope parameters (i.e., each feature vector to be accumulated) to obtain a total feature vector. It can be understood that the accumulation process is part of calculating the linear combination, which is used to generate the linear component of the predicted value in logistic regression.

[0080] The default intercept parameter refers to the bias term or intercept in the logistic regression model. As a constant term, the default intercept parameter does not depend on any input features, but is a fixed component of the model that is used to adjust the base level of the output results.

[0081] Specifically, it is necessary to first clarify which are the preset evasion features, which include but are not limited to total amount, toll, overload rate, vehicle model, total number of axles, etc., and can be set according to actual needs. Obtain the preset slope parameter (i.e., weight or coefficient) corresponding to each preset evasion feature and a preset intercept parameter from the evasion prediction model. Then, for each preset evasion feature, multiply the preset slope parameter corresponding to the preset evasion feature with the target feature vector of the target vehicle to generate a feature vector to be accumulated corresponding to the preset evasion feature. For example, if there is a preset evasion feature of "total amount", then multiply the target feature vector corresponding to the preset evasion feature of "total amount" by the preset slope parameter corresponding to the preset evasion feature of "total amount". Next, sum the feature vectors to be accumulated corresponding to all preset evasion features to obtain a total cumulative sum feature vector. This step actually constructs a part of the linear prediction function, which represents the comprehensive impact of all preset evasion features. And obtain the preset intercept parameter, and add the preset intercept parameter to the cumulative sum feature vector. Understandably, this preset intercept parameter represents the basic prediction level of the model even when all features do not exist. Finally, in order to convert the result of the linear combination into a probability form, an activation function is used to map the result obtained by adding the preset intercept parameter to the cumulative sum feature vector to the interval (0, 1) to obtain the probability of evasion.

[0082] Since the feature vector of the target vehicle is processed by the logistic regression model using the preset slope parameter (weight) and intercept parameter, the impact of each feature on the toll evasion behavior can be effectively quantified. This method can more accurately capture which features are highly correlated with toll evasion behavior, thereby improving the accuracy of toll evasion behavior prediction.

[0083] In one of the embodiments, a preset intercept parameter and a cumulative sum feature vector are added to obtain a linear feature vector; the linear feature vector is exponentially transformed to obtain a nonlinear feature vector; based on the nonlinear feature vector, the probability of the target vehicle evading tolls in the ETC lane is determined.

[0084] The linear feature vector is the result obtained by adding the preset intercept parameter to the cumulative feature vector. This process actually constructs the linear part of the logistic regression model, that is, the linear combination. It represents the comprehensive linear impact of all preset evasion features on whether the target vehicle will evade the fee.

[0085] The nonlinear feature vector refers to the output after exponential transformation. The "nonlinear feature vector" in this application is actually the result obtained after exponential transformation of the linear feature vector, the purpose of which is to map it to a probability space.

[0086] Specifically, the preset intercept parameter and the cumulative sum feature vector are first input, and the preset intercept parameter is added to the cumulative sum feature vector to obtain a linear combination, i.e., a linear feature vector. Then, an activation function (Sigmoid Function) is applied to the obtained linear combination to calculate the probability that the target vehicle has evaded tolls when passing through the ETC lane.

[0087] Optionally, according to the fare evasion prediction model, the calculation formula of the fare evasion probability P (y=1|x) is:

[0088] P(y=1|x)=1 / 1+

[0089] Among them, A is the preset intercept parameter, B is the preset slope parameter corresponding to each preset evasion feature, and xi is the target feature vector corresponding to each preset evasion feature.

[0090] Since a linear feature vector is obtained by adding the preset intercept parameter and the cumulative feature vector. This step allows each feature to have different degrees of influence on the final prediction result according to its importance (determined by the preset slope parameter). The linear feature vector is exponentially transformed, introducing nonlinear elements. This step allows the model to capture more complex, nonlinear relationships, thereby improving the model's expressiveness and enabling it to better fit the potential patterns in the data. Based on the nonlinear feature vector, the probability of the target vehicle evading tolls when passing through the ETC lane is calculated using logistic regression or similar methods. The output is a probability value between 0 and 1, and the results in this form are easy to understand and interpret.

[0091] In one of the embodiments, a preset probability threshold is obtained, and the probability that the target vehicle has evaded tolls in the ETC lane is compared with the preset probability threshold; when the probability that the target vehicle has evaded tolls in the ETC lane is greater than or equal to the preset probability threshold, the evasion behavior of the target vehicle in the ETC lane is classified based on the target feature vector of the target vehicle for each preset evasion feature.

[0092] The preset probability threshold refers to a critical value pre-set in the ETC toll inspection method. This critical value is used to distinguish whether a vehicle is likely to be involved in toll evasion. Specifically, it is a value between 0 and 1, representing the lower limit of the probability that the system determines that a vehicle is a toll evasion vehicle.

[0093] Specifically, first, a preset probability threshold is read from the system configuration file or other setting source. Optionally, the preset probability threshold can be a critical value determined in advance based on experience and data analysis, which is used to distinguish whether the target vehicle may be involved in toll evasion. Then compare whether the probability of the target vehicle's toll evasion behavior calculated in the above steps is greater than or equal to the preset probability threshold. Furthermore, if the probability of the target vehicle's toll evasion behavior is greater than or equal to the preset probability threshold, proceed to the next step; otherwise, end the process and consider the vehicle to be a normal vehicle. Finally, when the probability of the target vehicle's toll evasion behavior is greater than or equal to the preset probability threshold, the toll evasion behavior classification model receives the target feature vector as input, and the specific toll evasion behavior of the target vehicle is used as the classification result.

[0094] The reasonable setting of the preset probability threshold can help reduce the risk of misjudgment, ensuring that neither the real evaders will be let go nor unnecessary interference will be caused to normal users. And for those vehicles that are considered to have a high probability of evading fees, their evasion types are specifically classified (such as counterfeit, cheating, defective, and illegal). This helps to deeply understand the evasion model and formulate more targeted preventive measures.

[0095] In one of the embodiments, a first weight value and a first bias value corresponding to the evasion behavior classification category are obtained; for each preset evasion feature, based on the first weight value and the first bias value, the target feature vector for the preset evasion feature is converted into an activation value corresponding to the evasion behavior classification category; based on the activation value corresponding to the evasion behavior classification category, the evasion behavior probability distribution feature vector for the preset evasion feature is obtained; and according to the evasion behavior probability distribution feature vector corresponding to each preset evasion feature, the evasion behavior of the target vehicle in the ETC lane is classified.

[0096] The evasion behavior classification category refers to the classification of different types of evasion behaviors that may occur on the ETC lane. For example, counterfeiting, cheating, defect, and violation. Each category represents a specific type of evasion behavior pattern. In this application, it is used to further subdivide suspected evasion vehicles into specific evasion types so that more targeted management measures can be taken.

[0097] The first weight value refers to the initial weight setting related to the connection between the input layer and the hidden layer or directly to the output layer in the fare evasion behavior classification model. It can be understood that the first weight value determines the influence of each preset fare evasion feature on the final classification result.

[0098] The first bias value is another important parameter in the evasion behavior classification model. It refers to the bias term applied to the input layer or the first hidden layer in the evasion behavior classification model, which allows the model to activate certain neurons even when all inputs are zero. It greatly increases the flexibility of the model, allowing non-zero outputs to be generated even in the absence of input signals, which helps improve the model's fitting ability.

[0099] The activation value refers to the result obtained after applying the activation function. It can be understood that the activation function is used to introduce nonlinear elements so that the model can capture more complex patterns. In the present application, for each preset evasion feature, based on the first weight value and the first bias value, the activation value corresponding to the target feature vector is calculated. This step usually occurs at each layer of the neural network as part of the conversion from input to output.

[0100] The fare evasion behavior probability distribution feature vector refers to the vector representing the probability distribution of the target vehicle belonging to each fare evasion behavior classification category. Each element corresponds to a specific fare evasion behavior classification category and reflects the probability of the target vehicle evading the fare under this category.

[0101] Specifically, first, the first weight value and the first bias value corresponding to each evasion behavior classification category that have been pre-trained are extracted or loaded from the trained evasion behavior classification model. It can be understood that these parameters are learned through a large amount of historical data during the model training process and are used to describe the relationship between the feature and the evasion type. Each evasion behavior classification category corresponds to a weight matrix and a bias vector. Then, for each preset evasion feature, the formula Z=W is used. i × i +b i Compute the linear combination Z where x i is the target feature vector, W i and b i are the weight matrix and bias vector corresponding to the feature, respectively. Next, the activation function is applied to convert the linear combination Z into an activation value a. Each preset evasion feature will correspond to an activation value a. Further, the softmax function () or other appropriate normalization methods are used to convert the activation value a into a vector P representing the probability distribution of each evasion behavior classification category (i.e., the evasion behavior probability distribution feature vector). It can be understood that this step ensures that the sum of all probabilities is equal to 1, and each element is between 0 and 1, reflecting the probability that the target vehicle belongs to a specific evasion type. Finally, the probability distribution feature vector P of all preset evasion features is comprehensively considered, and the evasion behavior classification category with the highest probability is selected as the final classification result.

[0102] Since judgments can be made on multiple feature dimensions through comprehensive evaluation of the probability distribution feature vector of evasion behavior, the risk of misjudgment caused by a single feature is reduced, and the reliability of decision-making is improved. This step can not only identify whether evasion behavior has occurred, but also further subdivide it into specific evasion types (such as counterfeiting, cheating, defects, and violations). This comprehensive classification capability helps to deeply understand the evasion model and provide support for the formulation of targeted preventive measures.

[0103] In one of the embodiments, a plurality of candidate evasion features related to the ETC lane are obtained; for each candidate evasion feature, the candidate evasion feature is output to an evasion decision tree, the evasion decision tree including at least one evasion node; the impurity reduction amount produced by the splitting of the candidate evasion feature when passing through each evasion node in the evasion decision tree is determined; based on the impurity reduction amount produced by the splitting at each evasion node, the average impurity reduction amount of the candidate evasion feature is determined; based on the average impurity reduction amount corresponding to each of the plurality of candidate evasion features, a plurality of preset evasion features are screened out from the plurality of candidate evasion features.

[0104] The candidate toll evasion features refer to various feature variables that may be related to toll evasion behavior on the ETC lane. These features may include but are not limited to vehicle type, total weight, toll amount, overload rate, entry and exit time, etc.

[0105] The fare evasion decision tree is a machine learning model that segments a data set through a series of conditional judgments (nodes) and ultimately assigns samples to different categories. In this application, the fare evasion decision tree is specifically used to analyze and classify fare evasion behavior.

[0106] The evasion node refers to the internal node in the evasion decision tree. Each node corresponds to a specific candidate evasion feature and performs data segmentation based on the value of the feature. In this application, each evasion node divides the data entering the node into two subsets by applying a condition (such as a threshold of a certain feature), and then recursively builds a more fine-grained subtree until a leaf node is reached.

[0107] The reduction in impurity refers to the measure of how much the "chaos" (impurity) of the data distribution is reduced after a candidate evasion feature is used to split at a certain node in the evasion decision tree.

[0108] The average impurity reduction refers to the average value of the impurity reduction of a candidate evasion feature at all relevant nodes in the entire evasion decision tree. It is one of the key indicators for evaluating the importance of the feature. By calculating the average impurity reduction for each candidate evasion feature, the impact of each feature on the classification effect can be quantified, helping to screen out the most predictive preset evasion features.

[0109] Specifically, firstly, all the features that may be related to toll evasion in the ETC system are taken as candidate toll evasion features, which may include but are not limited to vehicle type, total weight, toll amount, overload rate, import and export time, etc.

[0110] Then, for each candidate evasion feature, an existing evasion decision tree model is constructed or used. The model includes at least one evasion node (i.e., root node) to evaluate the effect of a specific feature in distinguishing evasion behaviors. And each evasion node corresponds to a conditional test to determine how the data is split into different child nodes. It can be understood how each candidate evasion feature performs in the evasion decision tree, especially how it affects the data distribution.

[0111] Next, for each candidate evasion feature, the impurity change before and after the split is calculated at each evasion node in the evasion decision tree. The specific calculation formula is as follows:

[0112]

[0113] Where I(N) is the original impurity of node N, P After splitting, it reaches each child node N The data ratio, I (N ) is the child node N The reduction in impurity.

[0114] Finally, for each candidate evasion feature, the average of its impurity reduction on all related evasion nodes is calculated. This step aims to quantify the importance of each feature in the entire decision tree. Furthermore, all candidate evasion features are ranked according to the size of the average impurity reduction. Several features with the highest average impurity reduction are selected as the final preset evasion features. These features are considered to be the key features that can most effectively distinguish evasion behavior.

[0115] By calculating the reduction in impurity of each candidate evasion feature when each node in the evasion decision tree is split, and further calculating the average impurity reduction, the impact of each feature on the classification result can be quantified. This method is based on data and statistical analysis, ensuring the scientificity and objectivity of the feature selection process. And the decision tree model is used for feature evaluation, which has a clear structure and is easy to explain. The conditional test on each node intuitively shows how the feature affects the classification result, so that both managers and technicians can understand how the model works. Finally, by screening out the most important preset evasion features, the interference of irrelevant or redundant features on model training is reduced, thereby improving the overall performance of the classification model, including indicators such as accuracy and recall.

[0116] In one embodiment, reference Figure 3 The processing flow of the ETC toll inspection system mainly includes the following stages:

[0117] Data collection and analysis phase: First, we start with the data stored in the highway network toll collection system. We conduct preliminary collection and analysis of these data to identify the classification of toll evasion behaviors.

[0118] Abnormal vehicle feature analysis stage: Based on data collection, the features of abnormal vehicles are further analyzed, including multi-dimensional information such as overall features, time features, spatial features, and toll features.

[0119] Data preprocessing phase: Data preprocessing is a key step to ensure the accuracy of subsequent model training. It includes data normalization (making data from different sources comparable), data discretization (converting continuous variables into categorical variables), and data set sampling (selecting representative samples for modeling).

[0120] Determine the toll evasion vehicle dataset stage: Through the above steps, the dataset used to train and test the toll evasion vehicle identification model is finally determined.

[0121] In the training set of toll evasion vehicle data, we use the random forest algorithm to select features and find out the key factors that affect toll evasion. Based on the selected features, we use the logistic regression model to predict the toll evasion probability. Finally, we use the neural network model to classify the toll evasion types.

[0122] Model verification and analysis phase: Verify the established model, evaluate its performance in practical applications, and make necessary adjustments and optimizations.

[0123] Toll evasion vehicle identification stage: Apply the trained and verified model to actual scenarios to achieve effective identification of toll evasion vehicles.

[0124] High-speed ETC business system stage: Ultimately, the identification results are fed back to the high-speed ETC business system to help the management department take timely measures to deal with toll evasion and maintain normal traffic order.

[0125] Through the entire process from raw data to final application in the above steps, the important role of big data analysis technology in improving the efficiency of ETC toll inspection has been realized.

[0126] In an exemplary embodiment, Figure 4 As shown, it includes steps 402 to 410. Among them:

[0127] Step 402, obtaining multiple candidate evasion features related to the ETC lane; for each candidate evasion feature, outputting the candidate evasion feature to an evasion decision tree, the evasion decision tree including at least one evasion node; determining the impurity reduction amount of the candidate evasion feature when it passes through each evasion node in the evasion decision tree; determining the average impurity reduction amount of the candidate evasion feature based on the impurity reduction amount generated by the splitting at each evasion node; screening multiple preset evasion features from the multiple candidate evasion features based on the average impurity reduction amount corresponding to each of the multiple candidate evasion features; determining the initial feature vector of the target vehicle for each preset evasion feature;

[0128] Step 404, obtaining preset evasion weights of respective preset evasion features;

[0129] Step 406, determining a target feature vector of the target vehicle for each preset fare evasion feature according to the preset fare evasion weight and the initial feature vector of the target vehicle for each preset fare evasion feature;

[0130] Step 408, obtaining a preset slope parameter corresponding to each preset evasion feature; for each preset evasion feature, multiplying the preset slope parameter of the preset evasion feature with the target feature vector to obtain a feature vector to be accumulated corresponding to the preset evasion feature; accumulating the feature vectors to be accumulated corresponding to each preset evasion feature to obtain an accumulated feature vector; obtaining a preset intercept parameter, and adding the preset intercept parameter to the accumulated feature vector to obtain a linear feature vector; performing an exponential transformation on the linear feature vector to obtain a nonlinear feature vector; based on the nonlinear feature vector, determining the probability that the target vehicle has evasion behavior in the ETC lane;

[0131] Step 410, obtain a preset probability threshold, and compare the probability that the target vehicle has evaded the fee in the ETC lane with the preset probability threshold; when the probability that the target vehicle has evaded the fee in the ETC lane is greater than or equal to the preset probability threshold, obtain a first weight value and a first bias value corresponding to the evasion behavior classification category; for each preset evasion feature, based on the first weight value and the first bias value, convert the target feature vector for the preset evasion feature into an activation value corresponding to the evasion behavior classification category; based on the activation value corresponding to the evasion behavior classification category, obtain the evasion behavior probability distribution feature vector for the preset evasion feature; and classify the evasion behavior of the target vehicle in the ETC lane according to the evasion behavior probability distribution feature vector corresponding to each preset evasion feature.

[0132] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0133] Based on the same inventive concept, the embodiment of the present application also provides a fare evasion behavior inspection device for implementing the fare evasion behavior inspection method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more fare evasion behavior inspection device embodiments provided below can refer to the limitations of the fare evasion behavior inspection method above, and will not be repeated here.

[0134] In an exemplary embodiment, Figure 5 As shown, a fee evasion behavior inspection device 500 is provided, comprising: a first determination module 502, an acquisition module 504, a second determination module 506 and a third determination module 508, wherein:

[0135] The first determination module 502 is used to obtain a plurality of preset evasion features related to the ETC lane, and determine an initial feature vector of the target vehicle for each preset evasion feature;

[0136] An acquisition module 504 is used to acquire a preset evasion weight of each of a plurality of preset evasion features;

[0137] A second determination module 506 is used to determine a target feature vector of the target vehicle for each preset fare evasion feature according to a preset fare evasion weight and an initial feature vector of the target vehicle for each preset fare evasion feature;

[0138] The third determination module 508 is used to determine the probability that the target vehicle has evaded tolls in the ETC lane based on the target feature vector of the target vehicle for each preset evasion feature.

[0139] In one embodiment, the third determination module 508 is used to obtain a preset slope parameter corresponding to each preset evasion feature; for each preset evasion feature, multiply the preset slope parameter of the preset evasion feature and the target feature vector to obtain the feature vector to be accumulated corresponding to the preset evasion feature; accumulate the feature vectors to be accumulated corresponding to each preset evasion feature to obtain a cumulative feature vector; obtain a preset intercept parameter, and based on the preset intercept parameter and the cumulative feature vector, determine the probability that the target vehicle has evaded the fee in the ETC lane.

[0140] In one embodiment, the third determination module 508 is used to add the preset intercept parameter and the cumulative sum feature vector to obtain a linear feature vector; perform an exponential transformation on the linear feature vector to obtain a nonlinear feature vector; and determine the probability of the target vehicle evading tolls in the ETC lane based on the nonlinear feature vector.

[0141] In one of the embodiments, the toll evasion behavior inspection device also includes a classification module 510, which is used to obtain a preset probability threshold and compare the probability that the target vehicle has an toll evasion behavior in the ETC lane with the preset probability threshold; when the probability that the target vehicle has an toll evasion behavior in the ETC lane is greater than or equal to the preset probability threshold, the toll evasion behavior of the target vehicle in the ETC lane is classified based on the target feature vector of the target vehicle for each preset toll evasion feature.

[0142] In one embodiment, the classification module 510 is used to obtain a first weight value and a first bias value corresponding to the evasion behavior classification category; for each preset evasion feature, based on the first weight value and the first bias value, convert the target feature vector for the preset evasion feature into an activation value corresponding to the evasion behavior classification category; based on the activation value corresponding to the evasion behavior classification category, obtain the evasion behavior probability distribution feature vector for the preset evasion feature; according to the evasion behavior probability distribution feature vector corresponding to each preset evasion feature, classify the evasion behavior of the target vehicle in the ETC lane.

[0143] In one of the embodiments, the first determination module 502 is used to obtain multiple candidate evasion features related to the ETC lane; for each candidate evasion feature, output the candidate evasion feature to the evasion decision tree, the evasion decision tree includes at least one evasion node; determine the impurity reduction amount caused by the splitting of the candidate evasion feature when passing through each evasion node in the evasion decision tree; based on the impurity reduction amount caused by the splitting at each evasion node, determine the average impurity reduction amount of the candidate evasion feature; based on the average impurity reduction amount corresponding to each of the multiple candidate evasion features, screen out multiple preset evasion features from the multiple candidate evasion features.

[0144] In another embodiment, if Figure 6 As shown, Figure 6 The structural block diagram of the evasion behavior inspection device in another embodiment includes: a first determination module 502, an acquisition module 504, a second determination module 506, a third determination module 508 and a classification module 510, wherein: the classification module 510 is used to obtain a first weight value and a first bias value corresponding to the evasion behavior classification category; for each preset evasion feature, based on the first weight value and the first bias value, the target feature vector of the preset evasion feature is converted into an activation value corresponding to the evasion behavior classification category; based on the activation value corresponding to the evasion behavior classification category, the evasion behavior probability distribution feature vector of the preset evasion feature is obtained; according to the evasion behavior classification category corresponding to each preset evasion feature, A behavior probability distribution feature vector is used to classify the evasion behavior of the target vehicle in the ETC lane, and multiple candidate evasion features related to the ETC lane are obtained; for each candidate evasion feature, the candidate evasion feature is output to the evasion decision tree, and the evasion decision tree includes at least one evasion node; the impurity reduction amount generated by the splitting of the candidate evasion feature when passing through each evasion node in the evasion decision tree is determined; based on the impurity reduction amount generated by the splitting when each evasion node is split, the average impurity reduction amount of the candidate evasion feature is determined; based on the average impurity reduction amount corresponding to each of the multiple candidate evasion features, multiple preset evasion features are screened out from the multiple candidate evasion features.

[0145] Each module in the above-mentioned fee evasion behavior inspection device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0146] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 7 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data related to the evasion behavior audit. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for evasion behavior audit is implemented.

[0147] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0148] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0149] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0150] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0151] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0152] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0153] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0154] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for investigating fee evasion, characterized in that: The method comprises: Acquire multiple preset toll evasion features related to the ETC lane, and determine an initial feature vector of the target vehicle for each of the preset toll evasion features; Obtaining a preset fee evasion weight for each of the plurality of preset fee evasion features; Determining a target feature vector of the target vehicle for each of the preset fare evasion features according to the preset fare evasion weight and the initial feature vector of the target vehicle for each of the preset fare evasion features; Based on the target feature vector of the target vehicle for each of the preset toll evasion features, the probability that the target vehicle has toll evasion behavior when passing through the ETC lane is determined.

2. The method according to claim 1, characterized in that The determining, based on the target feature vector of the target vehicle for each of the preset toll evasion features, the probability that the target vehicle has toll evasion behavior in the ETC lane includes: Obtaining a preset slope parameter corresponding to each of the preset fee evasion features; For each of the preset evasion features, a preset slope parameter of the preset evasion feature and a target feature vector are multiplied to obtain a feature vector to be accumulated corresponding to the preset evasion feature; Accumulate the feature vectors to be accumulated corresponding to each of the preset fee evasion features to obtain an accumulated feature vector; A preset intercept parameter is obtained, and based on the preset intercept parameter and the cumulative sum feature vector, a probability that the target vehicle has evaded tolls in the ETC lane is determined.

3. The method according to claim 2, characterized in that The determining, based on the preset intercept parameter and the cumulative sum feature vector, the probability that the target vehicle has a toll evasion behavior in the ETC lane includes: Adding the preset intercept parameter and the accumulated feature vector to obtain a linear feature vector; Performing exponential transformation on the linear feature vector to obtain a nonlinear feature vector; Based on the nonlinear characteristic vector, the probability that the target vehicle has engaged in toll evasion behavior while passing through the ETC lane is determined.

4. The method according to claim 1, characterized in that The method further comprises: Obtaining a preset probability threshold, and comparing the probability that the target vehicle has evaded tolls in the ETC lane with the preset probability threshold; When the probability that the target vehicle has evaded tolls when passing through the ETC lane is greater than or equal to the preset probability threshold, the evasion behavior of the target vehicle when passing through the ETC lane is classified based on the target feature vector of the target vehicle for each of the preset evasion features.

5. The method according to claim 4, characterized in that The classifying the toll evasion behavior of the target vehicle in the ETC lane based on the target feature vector of the target vehicle for each preset toll evasion feature includes: Acquire a first weight value and a first bias value corresponding to the classification category of the fee evasion behavior; For each of the preset evasion features, based on the first weight value and the first bias value, converting the target feature vector for the preset evasion feature into an activation value corresponding to the evasion behavior classification category; Based on the activation value corresponding to the classification category of the evasion behavior, a probability distribution feature vector of the evasion behavior for the preset evasion feature is obtained; According to the evasion behavior probability distribution feature vector corresponding to each of the preset evasion features, the evasion behavior of the target vehicle in the ETC lane is classified.

6. The method according to any one of claims 1 to 5, characterized in that: The obtaining of multiple preset toll evasion features related to passing through the ETC lane includes: Acquire multiple candidate toll evasion features related to the ETC lane; For each of the candidate evasion features, outputting the candidate evasion features to a evasion decision tree, wherein the evasion decision tree includes at least one evasion node; Determine the amount of impurity reduction caused by splitting the candidate evasion feature when passing through each evasion node in the evasion decision tree; Determine an average impurity reduction amount for the candidate evasion features based on the impurity reduction amount produced by the splitting of each of the evasion nodes; Based on the average impurity reduction amount corresponding to each of the plurality of candidate evasion features, a plurality of preset evasion features are screened out from the plurality of candidate evasion features.

7. A device for checking fee evasion, characterized in that: The device comprises: A first determination module is used to obtain a plurality of preset evasion features related to the ETC lane, and determine an initial feature vector of the target vehicle for each of the preset evasion features; An acquisition module, configured to acquire a preset evasion weight of each of the plurality of preset evasion features; A second determination module is used to determine a target feature vector of the target vehicle for each of the preset fare evasion features according to the preset fare evasion weight and the initial feature vector of the target vehicle for each of the preset fare evasion features; The third determination module is used to determine the probability that the target vehicle has evaded tolls when passing through the ETC lane based on the target feature vector of the target vehicle for each of the preset evasion features.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images