User charging behavior modeling method and system

Abstract

The invention provides a user charging behavior modeling method and system, and the method comprises the steps: obtaining the global charging behavior data of a user, carrying out the analysis of the global charging behavior data of the user through a clustering method, and extracting the global charging behavior preference characteristics of the user; on the basis of global charging behavior preference characteristics, key factors influencing user behavior categories are analyzed by using information gain, a user behavior preference classification model is constructed, and charging behavior preferences of the user in each scene are determined; and obtaining historical charging behavior data of the user, and based on the historical charging behavior data and the key factors, utilizing the trained user charging behavior model to predict the user charging behavior of the target prediction day. By analyzing and modeling the charging behaviors of the user in different scenes, the efficiency of the power system and the charging experience of the user are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power systems and smart grids, and specifically relates to a user charging behavior modeling method and system. Background Art

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

[0003] With the transformation of the global energy structure and the rapid popularization of electric vehicles (EVs), the power system is facing unprecedented challenges and opportunities. The randomness and uncertainty of user charging behavior have a significant impact on the stability and efficiency of the power grid, especially during peak hours, when the surge in charging demand may lead to drastic fluctuations in the grid load. Traditional power system design mainly targets fixed loads and lacks effective management strategies for this emerging variable load. In other words, traditional power system design often does not fully consider the complexity of user charging behavior, which may lead to problems such as insufficient power supply or uneven grid load during peak hours.

[0004] The development of smart grid technology provides new tools for real-time monitoring, analysis and control of power systems, enabling power systems to better adapt to the access of distributed energy resources such as electric vehicle charging. However, user charging behavior is affected by many factors, including personal habits, availability of charging facilities, electricity price policies, environmental conditions, etc. The interaction of these factors makes user charging behavior highly complex and dynamic.

[0005] Existing modeling methods often rely on simplified assumptions and statistical models, which are difficult to accurately capture the diversity and dynamic changes of user behaviors. In addition, these methods usually lack a deep understanding of the motivations and preferences behind user behaviors. Summary of the invention

[0006] In order to solve the above problems, the present invention proposes a user charging behavior modeling method and system. The present invention analyzes and models the user's charging behavior in different scenarios to improve the efficiency of the power system and the user's charging experience; it integrates data analysis and machine learning technologies, including cluster analysis, decision trees and deep learning, to process and analyze large-scale user charging behavior data, which can extract valuable information from complex data and build accurate prediction models.

[0007] According to some embodiments, the present invention adopts the following technical solutions:

[0008] A user charging behavior modeling method includes the following steps:

[0009] Obtain the user's global charging behavior data, analyze the user's global charging behavior data using a clustering method, and extract the user's global charging behavior preference characteristics;

[0010] Based on the global charging behavior preference characteristics, the key factors affecting the user behavior category are analyzed using information gain, and a user behavior preference classification model is constructed to determine the user's charging behavior preference in each scenario.

[0011] The user's historical charging behavior data is obtained, and based on the historical charging behavior data and the key factors, the user's charging behavior on the target prediction day is predicted using the trained user charging behavior model.

[0012] As an optional implementation manner, the global charging behavior data includes some of arrival time, charging time, power-off time, departure time, charging cycle, charging load and charging amount.

[0013] As an optional implementation, a Gaussian mixture model is used to fit the user's global charging behavior data to determine the distribution of the data;

[0014] The charging cycles in the global charging behavior data are extracted using spectrum analysis, and low-frequency features are extracted as cycle feature data.

[0015] As an optional implementation method, the process of using information gain to analyze the key factors affecting user behavior categories and constructing a user behavior preference classification model includes:

[0016] Integrate other user charging attributes and user preference types to build a training data set;

[0017] Use information gain to calculate the information gain of other attributes relative to the user's preference type, and determine the key factors that affect the user's charging behavior preference;

[0018] Recursively calculate the information gain of other attribute information to the target attribute, build a decision tree model, and judge the user behavior preference of a single user in the target scenario.

[0019] As a further implementation manner, other charging attributes include charging season, whether it is a weekday, temperature and humidity.

[0020] As an optional implementation, user charging behavior models based on deep neural networks are constructed for different types of user charging behavior preferences, and when making predictions, the user charging behavior model that is compatible with the user charging behavior preference is selected.

[0021] As an optional implementation, the deep neural network is one of a long short-term memory neural network, a deep residual network, and a deep belief network.

[0022] As an optional implementation, the historical charging behavior data includes the return home time within the historical N days, whether charging information was charged within the historical N days, the return home time within the historical N days, whether charging was charged within the historical N days, the historical n charging times on the target prediction day, and the historical m departure times on the target prediction day, where N, n and m are integers and the value of N can be different.

[0023] As an optional implementation, the output of the user charging behavior model is the user charging probability, the user charging start time and the user leaving time.

[0024] A user charging behavior modeling system, comprising:

[0025] A global behavior preference feature extraction module is configured to obtain the user's global charging behavior data, analyze the user's global charging behavior data using a clustering method, and extract the user's global charging behavior preference features;

[0026] The scenario charging behavior preference determination module is configured to build a user behavior preference classification model based on the global charging behavior preference characteristics and use information gain to analyze the key factors affecting the user behavior category, so as to determine the user's charging behavior preference in each scenario;

[0027] The charging behavior prediction module is configured to obtain the user's historical charging behavior data, and based on the historical charging behavior data and the key factors, use the trained user charging behavior model to predict the user's charging behavior on the target prediction day.

[0028] A computer-readable storage medium is used to store computer instructions, and when the computer instructions are executed by a processor, the steps in the above method are completed.

[0029] An electronic device comprises a memory and a processor, and computer instructions stored in the memory and executed on the processor. When the computer instructions are executed by the processor, the steps in the above method are completed.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention can comprehensively consider personal habits, the availability of charging facilities, electricity price policies, environmental conditions and other factors, and accurately predict and analyze the modeling method of user charging behavior, which is of great significance for optimizing the allocation of power resources, improving the efficiency of power grid operation, enhancing user experience and promoting the healthy development of the electric vehicle industry. At the same time, studying the group and individual characteristics of user charging behavior is also of great significance for optimizing the allocation of power resources and reducing power grid load fluctuations.

[0032] The present invention combines unsupervised learning (such as cluster analysis) and supervised learning (such as deep neural networks) to conduct group and individual joint analysis and modeling of user charging behaviors, which helps to improve the accuracy and pertinence of the final prediction results.

[0033] The present invention uses GMM to perform cluster analysis, which can process various distributions in user behavior data and extract the charging preferences of group users, so that the model can adapt to complex data characteristics.

[0034] The present invention classifies user behavior preferences through a decision tree method, and can identify key factors that affect user charging behavior preferences, thereby achieving more accurate user behavior prediction; when constructing a decision tree model, information gain analysis is used to determine the key factors that affect user charging behavior preferences, which helps to improve the interpretability of the model and the accuracy of prediction.

[0035] The present invention uses a deep neural network to construct a user charging behavior model, which can predict whether a single user charges, the time to start charging, the time to leave, etc., thereby improving the accuracy and practicality of the prediction.

[0036] By combining decision tree classification and deep learning modeling, the present invention can build a personalized charging behavior prediction model for different types of user charging behavior preferences, thereby improving the personalization level of the prediction.

[0037] By predicting user charging behavior, the present invention can optimize the resource allocation of charging stations, reduce the user's charging waiting time, thereby improving the user's charging experience, and can provide support for load management of smart grids. By predicting user charging behavior, it helps to reduce the peak-to-valley difference in grid load and improve the stability and efficiency of the grid; it can also help power operators and policy makers better understand and manage electric vehicle charging behavior, thereby making more effective decisions.

[0038] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0040] Figure 1 It is a schematic diagram of the joint analysis and modeling process of user charging behavior groups and individuals;

[0041] Figure 2 It is a schematic diagram of the process of extracting user charging behavior preferences based on Gaussian mixture model;

[0042] Figure 3 This is a schematic diagram of the process of classifying user preferences based on a decision tree;

[0043] Figure 4 This is a schematic diagram of the process of predicting user behavior based on deep learning. DETAILED DESCRIPTION

[0044] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0045] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0046] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0047] In the absence of conflict, the embodiments in this application and the features in the embodiments may be combined with each other.

[0048] Embodiment 1

[0049] A user charging behavior modeling method includes the following steps:

[0050] Obtain the user's global charging behavior data, analyze the user's global charging behavior data using a clustering method, and extract the user's global charging behavior preference characteristics;

[0051] Based on the global charging behavior preference characteristics, the key factors affecting the user behavior category are analyzed using information gain, and a user behavior preference classification model is constructed to determine the user's charging behavior preference in each scenario.

[0052] The user's historical charging behavior data is obtained, and based on the historical charging behavior data and the key factors, the user's charging behavior on the target prediction day is predicted using the trained user charging behavior model.

[0053] A detailed introduction is given below.

[0054] like Figure 1As shown, the present invention aims at the fact that the charging preferences of group users often present the characteristics of diversity, while the charging behavior preferences of a single user present the characteristics of uncertainty. In order to realize the joint analysis and modeling of group and individual user charging behaviors, the unsupervised model (clustering) and the supervised model (deep neural network) are combined to realize the joint analysis and modeling of group and individual user charging behaviors. First, the clustering method is used to analyze the global user charging behavior data (such as home time, charging time, power off time, departure time, charging amount, etc.) to extract the global charging preferences of group users; on this basis, the key factors affecting the user behavior category (temperature, humidity, season, etc.) are analyzed based on information gain; finally, the deep learning method is used to construct the user individual charging behavior model.

[0055] First, a clustering method is used to analyze the global user charging behavior data. In this embodiment, a Gaussian mixture model is used to analyze the global user charging behavior data.

[0056] Gaussian Mixed Model (GMM) is a probabilistic clustering method that assumes that all data samples are generated by a mixture distribution composed of multiple mixed multivariate Gaussian distributions. In theory, GMM can fit any type of distribution and is used to solve clustering analysis when the data in the same set contains multiple different distributions (normal distribution, Bernoulli distribution, etc.).

[0057] like Figure 2 As shown in the figure, considering that the user's behavior data contains multiple distributions, GMM is used to implement cluster analysis of user charging behavior and extract the global charging preferences of group users. User charging behavior data includes charging time, arrival time, power outage time, departure time, charging load, charging cycle, etc. Among them, the charging cycle is extracted by spectrum analysis, and low-frequency features are extracted as cycle feature data.

[0058] A user has different charging behavior preferences in different scenarios due to factors such as season and weather. In order to classify user behavior preferences in different scenarios, information gain is used to determine the key factors affecting user charging behavior preferences, and a decision tree method is used to build a user behavior preference classification model.

[0059] like Figure 3 As shown, the specific steps are as follows:

[0060] Step 1: Construct a training data set by integrating the user's charging season, whether it is a weekday, temperature, humidity, etc. and the user's preference type;

[0061] Step 2: Use information gain to determine the information gain of other attributes (i.e. user charging season, whether it is a weekday, temperature, humidity, etc.) relative to the target attribute (user preference type) to determine the key factors affecting the user's charging behavior preference;

[0062] Step 3: Recursively calculate the information gain of other attribute information to the target attribute and build a decision tree model to determine the user behavior preference of a single user in a specific scenario.

[0063] According to different types of user charging behavior preferences, user charging behavior models based on deep neural networks are constructed. Figure 4 As shown in Figure 1, the model can predict whether a single user is charging, when to start charging, and when to leave. The deep neural network can use long short-term memory neural network, deep residual network, deep belief network, etc.

[0064] The inputs of the user charging behavior model of this embodiment are: the time of returning home within the past 7 days, whether the car was charged in the past 7 days (0 means not charged, 1 means charged), the time of returning home within the past 7 days, whether the car was charged in the past seven days (0 means not charged, 1 means charged), the historical n charging times on the target prediction day, the historical m departure times on the target prediction day, and the key influencing factors judged by information gain.

[0065] The outputs of the user charging behavior model of this embodiment are: user charging probability, user charging start time, and user leaving time.

[0066] To predict the charging behavior of individual users, the following steps can be used:

[0067] The user behavior preference classification model based on decision tree is used to determine the user's charging behavior preference type on the target forecast day;

[0068] The user charging behavior model corresponding to the charging behavior preference type is selected to predict the user's charging behavior on the target prediction day.

[0069] Embodiment 2

[0070] A user charging behavior modeling system, comprising:

[0071] A global behavior preference feature extraction module is configured to obtain the user's global charging behavior data, analyze the user's global charging behavior data using a clustering method, and extract the user's global charging behavior preference features;

[0072] The scenario charging behavior preference determination module is configured to build a user behavior preference classification model based on the global charging behavior preference characteristics and use information gain to analyze the key factors affecting the user behavior category, so as to determine the user's charging behavior preference in each scenario;

[0073] The charging behavior prediction module is configured to obtain the user's historical charging behavior data, and based on the historical charging behavior data and the key factors, use the trained user charging behavior model to predict the user's charging behavior on the target prediction day.

[0074] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0075] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0076] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0077] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0078] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principle of the present invention without creative labor shall be included in the protection scope of the present invention.

Claims

1. A user charging behavior modeling method, characterized in that: The following steps are involved: Obtain the user's global charging behavior data, analyze the user's global charging behavior data using a clustering method, and extract the user's global charging behavior preference characteristics; Based on the global charging behavior preference characteristics, the key factors affecting the user behavior category are analyzed using information gain, and a user behavior preference classification model is constructed to determine the user's charging behavior preference in each scenario. The user's historical charging behavior data is obtained, and based on the historical charging behavior data and the key factors, the user's charging behavior on the target prediction day is predicted using the trained user charging behavior model.

2. A user charging behavior modeling method as claimed in claim 1, characterized in that: The global charging behavior data includes some of arrival time, charging time, power-off time, departure time, charging cycle, charging load and charging amount.

3. A user charging behavior modeling method as claimed in claim 1, characterized in that: Use the Gaussian mixture model to fit the user's global charging behavior data and determine the distribution of the data; The charging cycles in the global charging behavior data are extracted using spectrum analysis, and low-frequency features are extracted as cycle feature data.

4. A user charging behavior modeling method as claimed in claim 1, characterized in that: The process of using information gain to analyze the key factors that affect user behavior categories and constructing a user behavior preference classification model includes: Integrate other user charging attributes and user preference types to build a training data set; Use information gain to calculate the information gain of other attributes relative to the user's preference type, and determine the key factors that affect the user's charging behavior preference; Recursively calculate the information gain of other attribute information to the target attribute, build a decision tree model, and judge the user behavior preference of a single user in the target scenario.

5. A user charging behavior modeling method as claimed in claim 4, characterized in that: Other charging attributes include charging season, whether it is a weekday, temperature and humidity.

6. A user charging behavior modeling method as claimed in claim 1, characterized in that: User charging behavior models based on deep neural networks are constructed for different types of user charging behavior preferences. When making predictions, the user charging behavior model that matches the user's charging behavior preferences is selected.

7. A user charging behavior modeling method as claimed in claim 1, characterized in that: The deep neural network is one of a long short-term memory neural network, a deep residual network and a deep belief network.

8. A user charging behavior modeling method as claimed in claim 1, characterized in that: The historical charging behavior data includes the return home time within the past N days, whether the car was charged within the past N days, the historical n charging times on the target prediction day, and the historical m departure times on the target prediction day. N, n, and m are integers, and the value of N can be different.

9. A user charging behavior modeling method as claimed in claim 1, characterized in that: The outputs of the user charging behavior model are the user charging probability, the user charging start time and the user leaving time.

10. A user charging behavior modeling system, characterized in that: include: A global behavior preference feature extraction module is configured to obtain the user's global charging behavior data, analyze the user's global charging behavior data using a clustering method, and extract the user's global charging behavior preference features; The scenario charging behavior preference determination module is configured to build a user behavior preference classification model based on the global charging behavior preference characteristics and use information gain to analyze the key factors affecting the user behavior category, so as to determine the user's charging behavior preference in each scenario; The charging behavior prediction module is configured to obtain the user's historical charging behavior data, and based on the historical charging behavior data and the key factors, use the trained user charging behavior model to predict the user's charging behavior on the target prediction day.

Images