Charging selection probability prediction method and system for expressway EV users

Abstract

The invention discloses a charging selection probability prediction method and system for highway EV users, and relates to the technical field of charging behavior prediction. The method comprises the following steps: obtaining personal attribute data and charging state data to be predicted, and carrying out preprocessing operation; establishing a charging behavior prediction model by using a single-class support vector machine, training the charging behavior prediction model by using a multi-dimensional training sample set, designing a target function according to classification of user charging behaviors, solving the target function to obtain a decision value representing user charging behavior characteristics, and calculating the charging behavior of the user according to the decision value. Deriving the decision value based on a density interval division method; and performing charging selection probability prediction on the personal attribute data and the charging state data by using the trained charging behavior prediction model. According to the invention, accurate charging selection behavior prediction can be carried out according to the historical data and the real-time state of the EV user, and powerful data support is provided for the intelligent service of the electric vehicle.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of charging behavior prediction, in particular to a charging selection probability prediction method and system for highway EV users. BACKGROUND

[0002] The statements in this section merely provide background information related to the present application and do not necessarily constitute prior art.

[0003] In recent years, the new energy vehicle industry has shown explosive growth, and as an important transportation network connecting cities and supporting long-distance travel, the layout of the charging infrastructure of the highway is directly related to the convenience and user experience of long-distance travel of electric vehicles, and further affects the further promotion of new energy vehicles. However, in the field of research on charging selection behavior of electric vehicle (EV) users, there are still some deficiencies in existing research, which is difficult to meet the needs of accurate and efficient charging selection probability modeling in the highway scenario.

[0004] First, the data source has certain limitations. At present, a large number of studies rely on online questionnaire research, which designs simulated charging or non-charging scenarios for participants to make choices to obtain state information of charging decisions. However, this method has inherent defects: first, it is difficult to recover, and the state information obtained depends on the charging memory of the respondents. Generally, only interval type option data of various charging state parameters are collected, not real-time and specific numerical state data of users; second, the subjectivity of data collection is very strong, and the answers of participants are easily affected by research situations, personal subjective cognitive bias and other factors, which is difficult to truly reflect the actual charging decision behavior. On the other hand, some studies use electric vehicle trajectory data sets for work. This kind of data belongs to unsupervised learning samples, which usually needs to use clustering methods such as K-means and hierarchical clustering for manual labeling. The core of these methods focuses on the similarity of data, and the subjective judgment of researchers is mixed in the labeling process, which makes it difficult for the model to accurately depict the influence of individual differences on charging selection.

[0005] Second, the adaptability of the model selection is insufficient. In the field of behavior prediction and behavior probability prediction, Logit and a series of derivative models have been widely used. However, the dependent variable of the Logit model is usually binary type, which can only output the decision result of "charging or not charging"; although the Logistic model can output the charging probability, when there is strong correlation between independent variables (such as charging station distance and travel time), it will cause the problem of unstable parameter estimation, and this kind of model is generally combined with clustering methods such as K-means, which further limits its application effect in complex scenarios. SUMMARY

[0006] In view of the deficiencies of the prior art, the purpose of the present application is to provide a charging selection probability prediction method and system for highway EV users, which can accurately predict the charging selection behavior of EV users according to historical data and real-time state of the EV users, and provide strong data support for intelligent services of electric vehicles.

[0007] In order to achieve the above-mentioned purpose, the present application is realized by the following technical scheme: The present application provides a charging selection probability prediction method for highway EV users in the first aspect, comprising the following steps: Obtain the personal attribute data and charging state data to be predicted, and perform preprocessing operation; A charging behavior prediction model is built by using a single-class support vector machine, and the charging behavior prediction model is trained by using a multi-dimensional training sample set, wherein a target function is designed according to the classification of the user charging behavior, the decision value representing the user charging behavior characteristics is obtained by solving the target function, and the charging selection probability of the user is obtained by deducing the decision value based on a density interval division method; The charging behavior prediction model after training is used to predict the charging selection probability of the personal attribute data and the charging state data.

[0008] Further, the personal attribute data includes gender and travel purpose, and the charging state data includes remaining mileage, power at the time of entering the highway, power at the time of starting charging and cruising range.

[0009] Further, the preprocessing operation is data cleaning of the personal attribute data and the charging state data, deletion of null value and abnormal value data, and one-hot encoding of the personal attribute data.

[0010] Further, the specific steps of designing a target function according to the classification of the user charging behavior and solving the target function to obtain the decision value representing the user charging behavior characteristics are as follows: A target function based on a hyperplane is constructed according to the classification of the user charging behavior, the inside of the hyperplane is a normal sample, i.e. a charging EV user, and the outside is an abnormal sample, i.e. a non-charging EV user; The target function is optimized by introducing a dual transformation and a kernel function; The optimized target function is solved to obtain the decision value.

[0011] Further, the specific steps of deducing the decision value based on the density interval division method are as follows: The decision value range is determined according to the decision value boundary; According to the density binning principle, the interval is divided according to the decision value range; Each interval is assigned a corresponding charging user sample probability.

[0012] Further, according to the density binning principle, the specific steps for interval division according to the decision value range are as follows: sorting the decision values in ascending order; dividing the sorted decision values into intervals with equal sample size according to the preset number of intervals.

[0013] The second aspect of the application provides a charging selection probability prediction system for highway EV users, comprising: a data acquisition module configured to acquire personal attribute data and charging state data to be predicted and perform preprocessing operations; a model training module configured to build a charging behavior prediction model using a single-class support vector machine and train the charging behavior prediction model using a multi-dimensional training sample set, wherein a target function is designed according to the classification of user charging behavior, the target function is solved to obtain a decision value representing the characteristics of user charging behavior, and the decision value is deduced based on a density interval division method to obtain the charging selection probability of the user; a probability prediction module configured to use the trained charging behavior prediction model to predict the charging selection probability of the personal attribute data and the charging state data.

[0014] The third aspect of the application provides a computer readable storage medium storing a computer program, the computer program being adapted to be loaded and executed by a processor to perform the steps of the charging selection probability prediction method for highway EV users as described in the first aspect of the application.

[0015] The fourth aspect of the application provides a computer device, comprising: a processor adapted to execute a computer program; a computer readable storage medium storing a computer program, the computer program being executed by the processor to implement the charging selection probability prediction method for highway EV users as described in the first aspect of the application.

[0016] The fifth aspect of the application provides a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to make the computer device perform the steps of the charging selection probability prediction method for highway EV users as described in the first aspect of the application.

[0017] The above one or more technical solutions have the following beneficial effects: The application discloses a charging selection probability prediction method and system for expressway EV users.

[0018] The application constructs a user charging selection problem as a hyperplane, introduces a dual problem transformation and a kernel function to simplify a target function, thereby reducing calculation complexity, and further introduces a density binning method based on a decision value to convert a user charging decision value obtained by the One-class SVM model into a charging probability, so that accurate prediction of the user charging selection probability is realized.

[0019] The advantages of the additional aspects of the application will be partially given in the following description, partially become obvious from the following description, or be learned by the practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0021] Figure 1 The flow chart of the charging selection probability prediction method for expressway EV users in the embodiment one of the application; Figure 2 The frequency distribution histogram of the charging state attribute data in the embodiment one of the application; Figure 3 The decision score schematic diagram of the One-class SVM model in the embodiment one of the application; Figure 4 The probability distribution schematic diagram of the One-class SVM model in the embodiment one of the application; Figure 5 The QQ diagram of the decision score in the embodiment one of the application. DETAILED DESCRIPTION

[0022] It should be noted that the following detailed description is exemplary in nature and is intended to provide further description of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0023] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments in accordance with the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, steps, operations, elements, components, and / or groups thereof, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof. The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0024] Embodiment one The embodiment one of the present application provides a charging selection probability prediction method for highway EV users, as shown in the following steps: Figure 1 S1: Obtain the personal attribute data and charging state data to be predicted, and perform preprocessing operation. In a specific embodiment, the personal attribute data includes gender and travel purpose, and the charging state data includes remaining mileage, power at highway entry, power at starting charging and cruising range. The preprocessing operation is to perform data cleaning on the personal attribute data and charging state data, delete null and abnormal value data, and perform One-Hot Encoding on the personal attribute data, map the classification value to an integer value, and then represent each integer value as a binary vector.

[0025] S2: Build a charging behavior prediction model using single-class support vector machine, and train the charging behavior prediction model using multi-dimensional training sample set.

[0026] S2.1: Obtain the multi-dimensional training sample set.

[0027]

[0028] ​In one specific implementation, this embodiment conducted on-site surveys in multiple service areas to obtain a training sample set. EV users' personal attributes and charging status data were recorded in detail through oral questioning, as shown in Table 1. The on-site surveys revealed that most service areas are equipped with intelligent charging equipment that can display the initial charge level, real-time charging power, and final charge level. Therefore, to ensure the reliability of the model input data, this embodiment combines the data collected from the questionnaires with the real-time data acquired by the charging equipment, selecting six variables as inputs to the charging choice behavior model: gender, travel purpose, remaining mileage, charge level when entering the highway, initial charge level, and driving range. After data preprocessing, removing null values, outliers, and other erroneous data, a total of 446 sample data points were obtained.

[0029] Table 1. Sample Questionnaire

[0030] It should be noted that both field survey data and real-time acquired data have been obtained with the user's consent and are used solely for the research purposes of this embodiment. Furthermore, to ensure the quantity and quality of the dataset, this embodiment will expand the training sample set using existing public datasets or data simulations.

[0031] As shown in Table 2, descriptive statistics were performed on the collected sample data. Indicators such as mean and standard deviation were used to present the central tendency, dispersion, and distribution characteristics of the data, intuitively revealing hidden underlying patterns (such as user charging habits and vehicle range distribution). Figure 2 As shown in Table 2. (Analysis of Table 2 and...) Figure 2 It can be seen that when entering the highway, the average battery level of users is 90.98% and the median is 100%, indicating that most people will fully charge their batteries to ensure highway travel. Only a few users have batteries as low as 25% due to emergencies or inconvenience in charging. Moreover, the data distribution is relatively concentrated without extreme anomalies. Users generally charge when the remaining battery level is around 35%. At the same time, the range of 5%-90% and the relatively flat data distribution show that there are differences in charging habits among different users.

[0032] Table 2. Descriptive statistics of charging status attribute data

[0033] S2.2: Construct a charging behavior prediction model using a single-class support vector machine.

[0034] In a specific embodiment, in combination with the characteristics of the sample collection of the highway service area, the embodiment finds that the proportion of vehicles that do not charge upon entering the service area is significantly less than the charging vehicles, and most of the vehicles that do not charge are to rest, eat and other activities in the service area, and a small number of vehicles will stay to answer questions. This makes the collected sample belong to single-class data, that is, only contains real data in the user charging scenario. The traditional behavior probability prediction model is limited by the data structure and is difficult to solve the prediction problem of single-class sample. Therefore, the embodiment innovatively uses a one-class support vector machine (One-class SVM) model to predict the charging selection of the highway EV user.

[0035] The one-class SVM is a variant of the support vector machine (SVM), and the core goal is anomaly detection or novelty detection in an unsupervised scenario, that is, only using a single class of normal samples to train the model, learning the feature distribution boundary of the normal samples, and then judging whether the new sample belongs to the normal class (out of the boundary is abnormal).

[0036] S2.3: Use a multi-dimensional training sample set to train the charging behavior prediction model.

[0037] S2.3.1: According to the classification design objective function of the user charging behavior, the decision value representing the characteristics of the user charging behavior is obtained by solving the objective function.

[0038] S2.3.1.1: According to the user charging behavior classification, a hyperplane-based objective function is constructed, and the inside of the hyperplane is the normal sample, that is, the charging EV user, and the outside is the abnormal sample, that is, the non-charging EV user.

[0039] In a specific embodiment, the training sample set of the highway EV user charging behavior is : .

[0040] Wherein, is the number of EV user samples, and in the embodiment , is the user sample index, is the th user sample, is a real set, is the feature dimension, and in the embodiment The core idea of One-Class SVM can be evolved in this embodiment as learning a hyperplane that "surrounds normal samples (charging user behavior)" in a six-dimensional feature space containing gender, travel purpose, remaining mileage, power at entering highway, power at starting charging, and cruising range. The inside of the hyperplane is normal samples, i.e., charging EV users, and the outside is abnormal samples, i.e., non-charging EV users. To achieve this goal, the original problem is defined as follows: (1).

[0041] wherein, represents the normal vector of the separating hyperplane in the high-dimensional highway EV user charging behavior feature space; represents the distance from the hyperplane to the origin, i.e., the radius that determines the normal area; a fault tolerance mechanism is set, is a slack variable, is the slack variable of the i-th user sample, and the embodiment allows a small number of charging user samples to fall outside the hyperplane by setting a fault tolerance mechanism; is a regularization parameter representing the upper limit of the proportion of charging EV user samples in the training set that are determined to be abnormal samples; is a kernel function mapping, which maps the original low-dimensional sample to a high-dimensional feature space to solve the non-linear separable problem.

[0042] S2.3.1.2: Optimize the objective function by introducing dual transformation and kernel function.

[0043] In a specific embodiment, considering the extremely high computational complexity of the high-dimensional feature space of highway EV user charging behavior, One-Class SVM uses dual transformation and kernel function for simplified calculation. According to the Lagrange multiplier method, the Lagrange function is constructed for the original problem and the extreme value is found, which proves that the optimal solution satisfies: (2).

[0044] wherein, is a dual variable, satisfying .

[0045] Substitute formula (2) into the original constraint of formula (1), and all calculations can be realized by kernel function . The commonly used kernel function is radial basis function (RBF): (3).

[0046] wherein, is a kernel parameter that controls the similarity range of sample mapping, and ​Respectively, sample users and feature center points Kernel function mapping.

[0047] S2.3.1.3: Solve the optimized objective function to obtain the decision value.

[0048] In a specific embodiment, the One-Class SVM determines whether the EV user sample is a charging EV user sample through the decision value. Substitute formula (2) into the hyperplane equation, and the decision value formula of a certain EV user sample is: (4).

[0049] If : a certain EV user sample falls inside the hyperplane, it is determined to be a charging EV user sample; If : a certain EV user sample falls outside the hyperplane, it is determined to be a non-charging EV user sample.

[0050] S2.3.2: Derive the decision value based on the density partition method to obtain the charging selection probability of the user.

[0051] In a specific embodiment, since the One-Class SVM only outputs the decision value, it needs to be mapped to the probability in the interval through a probability conversion, i.e., reflecting the confidence of the sample being a charging user sample. Density binning is a non-parametric method based on the density partition of the decision value. The core principle is that the probability distribution needs to simulate the density distribution of the decision value of the training data, and the derivation process is as follows: Since the One-Class SVM has no real label, density binning is based on two key assumptions: 1) The decision value is positively correlated with the charging probability of the EV user: the larger the decision value , the higher the probability of the sample being a normal sample, because corresponds to a charging EV user sample, and the larger the value, the farther away from the abnormal (non-charging) boundary; 2) Probability continuity matches density: the density of the decision value region can reflect the concentration degree of the sample distribution. In the dense region of the decision value, the probability change should be more gentle; in the sparse region of the decision value, the probability change can be more steep, avoiding the excessive influence of outliers on the probability.

[0052] S2.3.2.1: Determine the decision value range according to the decision value boundary.

[0053] In a specific embodiment, first calculate the decision value of all samples of the highway EV charging user , determine the minimum value of the decision value , maximum value , and decision boundary (charge / no charge threshold) .

[0054] S2.3.2.2: According to the principle of density binning, interval division is performed according to the decision value range.

[0055] In a specific embodiment, the decision value is sorted from small to large. The sorted decision value is divided into equal sample size intervals according to the preset interval number. Unlike the equal interval binning rule which divides the interval according to the fixed interval, the density binning divides the interval according to the density distribution of the decision value. The specific rule is as follows: the decision value of the highway EV charging user is sorted from small to large: ; according to the preset interval number , the sorted decision value is divided into equal sample size intervals, i.e. each interval contains or samples; for example: if , , each interval contains 100 samples; the purpose is to make the interval width smaller in the dense area of the decision value and larger in the sparse area, matching the density distribution. Determine the boundary of each interval: the left boundary of the first interval is , and the right boundary is , where is the sample number of each interval.

[0056] S2.3.2.3: Assign a corresponding charging user sample probability to each interval.

[0057] In a specific embodiment, a corresponding charging user sample probability is assigned to each interval, which needs to meet two constraints: 1. Monotonicity of probability: the larger the decision value of the interval, the higher the probability assigned (consistent with hypothesis 1); 2. Boundary calibration: the probability of the interval where the decision boundary is located needs to be calibrated to 0.5 (because is the critical value of normal and abnormal, the confidence is 50%).

[0058] The specific probability assignment formula is derived as follows: 1. Let the EV user decision boundary fall into the first interval, and the probability of the interval is ; 2. For the abnormal side interval of , the probability is assigned in a linear decreasing manner, and the abnormal side interval has Individual: (6).

[0059] where the probability of the minimum interval is close to 0, reflecting high abnormal confidence.

[0060] 3. For the normal side interval of , the probability is assigned in a linearly increasing manner, and the normal side interval has individuals: (7).

[0061] where the probability of the maximum interval is close to 1, reflecting high normal confidence.

[0062] S3: Use the trained charging behavior prediction model to predict the charging selection probability of individual attribute data and charging state data.

[0063] In a specific embodiment, for any test EV user sample , after calculating the decision value : determine which interval it falls into. Directly use the probability of the interval as the probability of the charging user sample evaluated by

[0064] .

[0065] In order to better illustrate the superiority of the method of the present embodiment, the following experimental verification is carried out: Use Python software to establish a One-Class SVM model, and the model output is the probability of each EV user selecting charging in the service area. The model independent variable input can be divided into two categories: Numerical features, i.e. charging state data: power when driving at high speed, power when charging, cruising range, remaining range.

[0066] Category features, i.e. individual attribute data: gender, travel purpose.

[0067] Category features such as gender and travel purpose cannot be directly input into the model and need to be One-Hot encoded. This requires mapping the classification value to an integer value first. Then, each integer value is represented as a binary vector.

[0068] The most important thing in the training process of the One-Class SVM model is the regularization parameter ​, which represents the upper bound of the proportion of the training set judged as "atypical charging samples", that is, individual data that do not behave like typical charging behavior in the charging user data set, in combination with the results of field research, the present embodiment is set to 0.05, and those that meet the "high power (80-100%) but very low remaining mileage (0-100 km)" condition can be classified as "atypical charging samples". The model learns the charging characteristics of these atypical samples to classify charging and non-charging users.

[0069] The model uses a density binning method based on decision score to convert the decision score of the highway EV user into a probability. Considering the sample size, the number of intervals is set to 6, and each bin is mapped to a uniform probability of [0.001, 0.999] (avoiding 0 / 1 extreme values) The present embodiment uses decision score, probability distribution, and quantile-quantile plot (QQ plot) analysis to analyze the model training results, as shown in Figure 3 , Figure 4 and Figure 5 . As can be seen from Figure 3 and Figure 4 , the decision score distribution of the sample is obviously right-biased, with most samples concentrated in the 0.2-1.2 interval, indicating that the model considers most EV user samples to be "normal charging" behavior; there is a significant peak near 0 (the highest bar), indicating that many EV user samples are "just on the border", which may be marginal normal charging samples or slightly abnormal. Figure 5 The X-axis of the X-axis represents the theoretical quantile of the standard normal distribution, and the Y-axis represents the actual quantile of the sample decision score. The red diagonal line indicates that the sample data should ideally fall on this line (i.e. completely conforming to the normal distribution), and the blue points represent the actual quantile of the sample. As can be seen from Figure 5 , most points are close to the red diagonal line, indicating that the decision score of the training set is close to the normal distribution as a whole. On the left side (negative region), the points deviate significantly from the red line, indicating that there are charging abnormalities or boundary samples that can be distinguished from normal samples, and on the right side Figure 5 , the data points are slightly higher than the red line, indicating that the high segment samples are more "extreme" than expected by the normal distribution, which also conforms to the characteristics of "few abnormal samples" in the anomaly detection model. Therefore, the One-Class SVM model established in the present embodiment successfully identifies a small number of abnormal samples (low decision score), and the "predicted probability" obtained by quantile mapping is highly positively correlated with the original decision score, and the model output has business interpretation.

[0070] Embodiment Two: The present embodiment provides a charging selection probability prediction system for highway EV users, comprising: a data acquisition module configured to acquire personal attribute data and charging state data to be predicted and perform a preprocessing operation; a model training module configured to construct a charging behavior prediction model by using a single-class support vector machine and train the charging behavior prediction model by using a multi-dimensional training sample set, wherein a target function is designed according to a classification of user charging behaviors, the target function is solved to obtain a decision value representing a charging behavior feature of a user, and the decision value is deduced based on a density partition interval method to obtain a charging selection probability of the user; a probability prediction module configured to perform charging selection probability prediction on the personal attribute data and the charging state data by using the trained charging behavior prediction model.

[0071] Embodiment Three: Embodiment Three of the present application provides a computer readable storage medium storing a computer program, the computer program being adapted to be loaded and executed by a processor to perform the steps in the charging selection probability prediction method for expressway EV users as described in Embodiment One of the present application.

[0072] Embodiment Four: Embodiment Four of the present application provides a computer device, which comprises: a processor adapted to execute a computer program; a computer readable storage medium storing a computer program, the computer program being executed by the processor to perform the steps in the charging selection probability prediction method for expressway EV users as described in Embodiment One of the present application.

[0073] Embodiment Five: Embodiment Five of the present application provides a computer program product or a computer program, which comprises computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to make the computer device perform the steps in the charging selection probability prediction method for expressway EV users as described in Embodiment One of the present application.

[0074] The steps involved in Embodiments Two, Three, Four and Five above correspond to Embodiment One, and the specific implementation can be referred to the relevant description part of Embodiment One.

[0075] Those skilled in the art can clearly understand that the units and algorithm steps of each example described in combination with the embodiments disclosed in the present application can be implemented by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application. In the above embodiments, all or part can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions according to the embodiments of the present application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted by a computer-readable storage medium. Computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired or wireless means. The computer-readable storage medium can be any available medium that the computer can access or a data processing device such as a server, data center, etc. that includes one or more available media sets. The available media can be magnetic media, optical media or semiconductor media, etc. The above description is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any skilled person in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A charging selection probability prediction method for highway EV users, characterized by, The method comprises the following steps: obtaining personal attribute data and charging state data to be predicted and performing preprocessing operations; building a charging behavior prediction model by using a single-class support vector machine, and training the charging behavior prediction model by using a multi-dimensional training sample set, wherein a target function is designed according to the classification of user charging behaviors, the target function is solved to obtain a decision value representing the characteristics of user charging behaviors, and the decision value is deduced based on a density interval division method to obtain the charging selection probability of the user; predicting the charging selection probability of the personal attribute data and the charging state data by using the trained charging behavior prediction model.

2. The highway EV user-oriented charging selection probability prediction method according to claim 1, wherein, The personal attribute data includes gender and travel purpose, and the charging state data includes remaining mileage, power at high-speed, power at the start of charging and cruising range.

3. The highway EV user-oriented charging selection probability prediction method according to claim 1, characterized by, The preprocessing operation is data cleaning of the personal attribute data and the charging state data, deletion of null and abnormal value data, and one-hot encoding of the personal attribute data.

4. The highway EV user-oriented charging selection probability prediction method of claim 1, wherein, The specific steps of designing a target function according to the classification of user charging behaviors, solving the target function to obtain a decision value representing the characteristics of user charging behaviors are as follows: constructing a target function based on a hyperplane according to the classification of user charging behaviors, the inside of the hyperplane being normal samples, i.e. charging EV users, and the outside being abnormal samples, i.e. non-charging EV users; optimizing the target function by introducing a dual transformation and a kernel function; solving the optimized target function to obtain a decision value.

5. The highway EV user-oriented charging selection probability prediction method of claim 1, wherein, The specific steps of deducing the decision value based on the density interval division method are as follows: determining the decision value range according to the decision value boundary; dividing the interval according to the decision value range according to the density binning principle; allocating the corresponding charging user sample probability to each interval.

6. The highway-oriented EV user's charging selection probability prediction method according to claim 5, wherein, The specific steps of dividing the interval according to the decision value range according to the density binning principle are as follows: sorting the decision values in ascending order; dividing the sorted decision values into intervals with equal sample sizes according to the preset number of intervals.

7. A highway EV user oriented charging selection probability prediction system characterized by, The method comprises the following steps: a data acquisition module configured to obtain personal attribute data and charging state data to be predicted and perform preprocessing operations; a model training module configured to build a charging behavior prediction model by using a single-class support vector machine, and train the charging behavior prediction model by using a multi-dimensional training sample set, wherein a target function is designed according to the classification of user charging behaviors, the target function is solved to obtain a decision value representing the characteristics of user charging behaviors, and the decision value is deduced based on a density interval division method to obtain the charging selection probability of the user; a probability prediction module configured to predict the charging selection probability of the personal attribute data and the charging state data by using the trained charging behavior prediction model.

8. A computer program product, characterised in that, The computer program product comprises a computer program, which, when executed by a processor, implements the charging selection probability prediction method for highway EV users according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, which is suitable for being loaded and executed by a processor to implement the charging selection probability prediction method for highway EV users according to any one of claims 1-6.

10. A computer device, comprising: The method comprises the following steps: a processor suitable for executing a computer program; A computer readable storage medium, in which a computer program is stored, the computer program, when executed by the processor, implements the highway EV user oriented charging selection probability prediction method according to any one of claims 1-6.

Images