Data analysis device and operating method thereof

Abstract

According to an embodiment disclosed in this document, a data analysis apparatus may include: an information acquisition unit for acquiring a plurality of battery data sets; and a controller that generates, for each of the battery datasets, feature data representing a correlation between at least two factors for analyzing a type of the battery dataset, extracts, for each feature data, target data associated with at least one parameter based on a characteristic of the feature data, and transmits the extracted target data to the battery dataset. And clustering the feature data based on the target data.

Description

Technical Field

[0001] Cross-reference to related applications

[0002] This application claims priority to Korean Patent Application No. 10-2023-0110181, filed on August 22, 2023, the entire contents of which are incorporated herein by reference for all purposes. Technical Field

[0003] The embodiments disclosed in this document relate to a data analysis apparatus and its operating method. Background Technology

[0004] Recently, research and development on rechargeable batteries have been actively pursued. Here, rechargeable batteries refer to batteries that can be recharged, and include all conventional Ni / Cd batteries, Ni / MH batteries, and more recently, lithium-ion batteries. Among rechargeable batteries, lithium-ion batteries have the advantage of a significantly higher energy density than conventional Ni / Cd and Ni / MH batteries. Furthermore, lithium-ion batteries can be manufactured small and lightweight enough to be used as power sources for mobile devices, and their application has recently expanded to electric vehicles, thus attracting attention as a next-generation energy storage medium.

[0005] Clustering techniques, which group similar data together, are commonly used to analyze the characteristics of battery data. However, the characteristics of battery data are influenced by various factors, and when clustering is performed considering the correlation of these factors, the combination of multiple factors leads to high data dimensionality and increased computational complexity. Therefore, a data analysis method that can improve the efficiency and performance of battery data clustering is needed. Summary of the Invention

[0006] Technical issues

[0007] The purpose of the embodiments disclosed in this document is to provide a data analysis apparatus and its operating method that can improve clustering results while reducing the computational complexity in the battery data clustering process.

[0008] The technical objectives of the embodiments disclosed in this document are not limited to those described above, and other objectives not described herein will be clearly understood by those skilled in the art from the following description.

[0009] Technical solution

[0010] According to the embodiments disclosed in this document, a data analysis apparatus may include: an information acquisition unit configured to acquire a plurality of battery datasets; and a controller configured to, for each battery dataset, generate feature data representing the correlation between at least two factors used to analyze the type of the battery dataset, extract target data associated with at least one parameter based on the characteristics of the feature data for each feature data, and cluster the feature data based on the target data.

[0011] According to an implementation, the feature data may include contour image data.

[0012] According to an implementation, the at least one parameter may include at least one of the following: the number of peak points in the contour image data, the width and length of each peak point, and the centroid coordinates.

[0013] According to an implementation, the controller can calculate the number of peak points for each feature data and group the feature data based on the number of peak points.

[0014] According to the implementation method, the controller can cluster the feature data belonging to each group based on the width, length and centroid coordinates of the peak point of each feature data.

[0015] According to the implementation method, the controller can classify the type of the battery dataset based on the clustering results.

[0016] According to an implementation, the controller generates the feature data based on the joint probability density function (joint PDF) of the at least two factors.

[0017] According to the embodiments disclosed in this document, an operation method of a data analysis apparatus may include the following steps: acquiring a plurality of battery datasets; for each battery dataset, generating feature data representing the correlation between at least two factors used to analyze the type of battery dataset; for each feature data, extracting target data associated with at least one parameter based on the characteristics of the feature data; and clustering the feature data based on the target data.

[0018] According to an implementation, the feature data may include contour image data, and the at least one parameter includes at least one of the number of peak points in the contour image data, the width of each peak point, the length of each peak point, and the centroid coordinates.

[0019] According to an implementation, the step of calculating the target data may include the following steps: calculating the number of peak points for each feature data; grouping the feature data based on the number of peak points; and extracting the target data based on the width, length, and centroid coordinates of the peak points for each feature data.

[0020] According to an implementation method, the step of clustering the feature data based on the target data may include: for each group, clustering the feature data belonging to each group based on the target data.

[0021] According to an implementation, the method may further include classifying the type of the battery dataset based on the clustering results.

[0022] Beneficial effects

[0023] The data analysis apparatus and its operating method according to the embodiments disclosed in this document can reduce computational load and complexity by extracting target data with a reduced data size.

[0024] Furthermore, the data analysis apparatus and its operating method according to the embodiments disclosed in this document can improve clustering performance by more effectively reflecting the characteristics of the clustering target through clustering based on the target data.

[0025] In addition, various effects that can be identified directly or indirectly through this document can be provided. Attached Figure Description

[0026] Figure 1 This is a diagram illustrating the configuration of a data analysis apparatus according to an embodiment disclosed in this document; Figure 2 This is a diagram illustrating an example of feature data in the form of a contour image according to an embodiment disclosed in this document; Figure 3a and Figure 3b This is a diagram illustrating an example of characteristic data based on the number of peak points according to an implementation disclosed in this document; Figure 4 This is a flowchart illustrating an operation method of a data analysis apparatus according to an embodiment disclosed in this document; Figure 5 This is a flowchart illustrating the process of extracting target data by a data analysis apparatus according to an embodiment disclosed in this document; and Figure 6 This is a diagram illustrating the hardware configuration of a computing system for performing an operation method of a data analysis apparatus according to an embodiment disclosed in this document. Detailed Implementation

[0027] In the following description, various embodiments of the invention will be described with reference to the accompanying drawings. However, this description is not intended to limit the invention to the specific embodiments described herein, and it should be construed as including various modifications, equivalents, and / or substitutions to the embodiments described herein.

[0028] In this document, the singular noun corresponding to an item may contain one or more instances of that item unless explicitly indicated otherwise in the context. In this document, phrases such as “A or B,” “at least one of A and B,” “at least one of A or B,” “A, B, or C,” “at least one of A, B, and C,” and “at least one of A, B, or C” may each individually refer to any of the listed items or any combination of such items. Terms such as “first” and “second” are used only to distinguish one component from another and do not limit these components in other respects (e.g., importance or order). In this document, when a component (e.g., the first component) is referred to as “connected” or “linked” to another component (e.g., the second component) with or without terms such as “functionally” or “communically,” it means that the first component may be connected to the second component directly (e.g., via a wired connection), wirelessly, or through a third component.

[0029] Each component (e.g., module or program) described in this document may include one or more instances. Depending on various implementations, one or more components or operations described herein may be omitted, or one or more other components or operations may be added. Alternatively or additionally, multiple components (e.g., modules or programs) may be integrated into a single component. In this case, the integrated component may perform one or more functions of each of the multiple components in the same or similar manner as the functions performed by corresponding components among the multiple components prior to integration. Depending on various implementations, operations performed by modules, programs, or other components may be performed sequentially, in parallel, iteratively, or heuristically, or one or more operations may be performed in a different order, omitted, or one or more other operations may be added.

[0030] The terms “module” or “unit” as used herein may cover a unit implemented in hardware, software, or firmware, and may be used interchangeably with terms such as logic, logic block, component, or circuit. A module may be a component assembled as a whole or one or more units performing one or more functions, which may constitute the smallest unit of a component or a part thereof. For example, according to one implementation, a module may be implemented in the form of an application-specific integrated circuit (ASIC).

[0031] The various implementations of this document can be implemented as software (e.g., a program or application) including one or more instructions stored on a machine-readable storage medium (e.g., memory). For example, a device's processor can invoke and execute at least one instruction stored on the storage medium. This enables the device to operate to perform at least one function according to at least one invoked instruction. The one or more instructions may include code generated by a compiler or code executable by an interpreter. The machine-readable storage medium may be provided in the form of a non-transitory storage medium. Here, the term "non-transitory" means that the storage medium is a tangible device and does not include signals (e.g., electromagnetic waves), and is used regardless of whether the data is permanently or temporarily stored.

[0032] Figure 1 This is a diagram illustrating the configuration of a data analysis apparatus according to an embodiment disclosed in this document.

[0033] Reference Figure 1 The data analysis device 100 may include an information acquisition unit 110 and a controller 120.

[0034] The data analysis device 100 can cluster battery datasets to classify the types of battery datasets. The data analysis device 100 can generate feature data from each battery dataset and extract target data smaller than the feature data. Therefore, the data analysis device 100 can reduce the computational complexity of the clustering process and produce clustering results that more clearly reflect the features.

[0035] Information acquisition unit 110 can acquire multiple battery datasets. Each battery dataset may include time-series data of battery data. For example, information acquisition unit 110 can obtain multiple battery datasets from batteries installed in various vehicles. As an example, information acquisition unit 110 can receive battery data from the battery management system (BMS) of the battery pack installed in each vehicle and from the on-board diagnostic (OBD) device equipped in the vehicle. In an embodiment, data analysis device 100 can acquire battery datasets from each vehicle and determine the personalization type of the vehicle driver by analyzing the type of each battery dataset.

[0036] The controller 120 can generate feature data representing the correlation between at least two factors, which are used to analyze the type of each battery dataset.

[0037] At least two factors can each be of a type used to analyze the battery dataset. As an example, each factor can include at least the initial state of charge (SOC) at the start of charging, depth of discharge (DoD) at charging, charging rate (C), starting temperature at charging, maximum temperature at charging, charging start time, charging start date within a week, start time of driving, driving distance, starting temperature of driving, maximum temperature of driving, driving DoD, distribution of driving speeds, discharge time, and end date of discharge within a week. Each factor is not limited to these and can include unrestricted factors that may affect battery condition, such as the duration of idle state and ambient temperature.

[0038] Controller 120 can analyze the correlation between at least two factors. According to one implementation, controller 120 can analyze the correlation between at least two factors based on various operations capable of generating correlations (e.g., covariance and joint probability) between factors in each dataset. According to another implementation, controller 120 can analyze the correlation between at least two factors based on the joint probability density (joint PDF) of at least two factors.

[0039] The controller 120 can generate feature data representing the correlation between at least two factors. For example, the controller 120 can display the feature data in various ways, such as graphs, images, and histograms, to represent the correlation between at least two factors.

[0040] According to an implementation, the feature data may include contour image data. The controller 120 can generate feature data by representing the correlation between at least two factors as a contour image. The controller 120 may also represent the feature data not only in contour image format but also in other graphical or image formats. In the following description, for ease of explanation, it is described that the feature data includes contour image data.

[0041] According to an implementation, controller 120 can generate feature data based on the joint probability density (joint PDF) of at least two factors. Controller 120 can represent the joint probability density of at least two factors as contour image data. For example, controller 120 can generate a contour image by connecting points along the same line with the same probability density.

[0042] The controller 120 can also extract target data associated with at least one parameter based on the characteristics of each feature data. The parameter based on the characteristics of the feature data can include various parameters that can be extracted or derived from the feature data. For example, in contour image data, at least one parameter can include the number of peak points, the width, length, and coordinates of the center of each peak point. Additionally, various parameters can be included, such as the number of contour lines, the extent of their distribution, etc.

[0043] In one implementation, the controller 120 can extract target data from the feature data. The feature data represents the correlation between at least two factors, and due to its large size, clustering the feature data increases computational cost and complexity. Therefore, the controller 120 can reduce the data size by extracting target data associated with the factors from the feature data. This reduces computational cost and complexity during the clustering process.

[0044] The controller 120 can cluster feature data based on target data. When clustering is performed on feature data with a large size, all data in the feature data are reflected in the clustering process, which may reduce the discriminative power of the data and decrease the clustering results. Therefore, the controller 120 can cluster target data with a smaller size to effectively reflect the characteristics of each battery dataset and achieve clearer distinction of characteristics within each cluster, thereby enhancing the classification results. Thus, it can solve the problem of mixed clustering that may occur when clustering with large data due to reduced discriminative power of the data.

[0045] Since the target data and feature data have a corresponding relationship, the controller 120 can cluster the feature data into multiple clusters according to the clustering results of the target data, so that the feature data corresponds to the clustering results of the target data.

[0046] According to one implementation, the controller 120 can calculate the number of peak points for each feature data. The controller 120 can calculate the number of peak points in a contour image of the feature data. For example, a peak point may refer to the point in the contour image where the data value (such as a probability density value) is highest. In another example, a peak point may refer to a contour line representing the highest data value.

[0047] According to one implementation, the controller 120 can group feature data based on the number of peak points. Grouping feature data based on the number of peak points enhances clustering results by using the same criteria to cluster the feature data. The controller 120 can group feature data based on the number of peak points because the data size of the target data may differ when the number of peak points varies. For example, the controller 120 can group feature data with the same number of peak points into a single group.

[0048] According to the implementation, the controller 120 can extract target data based on the width, length, and centroid coordinates of each peak point of each feature data. Here, the width of the peak point can refer to the x-axis width of the peak point in the contour image, the length can refer to the y-axis length of the peak point in the contour image, and the centroid coordinates can refer to the x-axis and y-axis coordinates of the peak center in the contour image. In order to perform clustering under a unified standard, the controller 120 can uniformly set the size of the contour image.

[0049] In this example, controller 120 can extract data with (4) The target data is the number of peak points. That is, the controller 120 can extract the width, length, center x-coordinate, and center y-coordinate of each peak. By extracting the target data in this way, the controller 120 can group the feature data based on the number of peak points, which changes the size of the data.

[0050] According to an implementation, controller 120 can cluster feature data based on target data for each group. For example, controller 120 can classify target data with similar characteristics into the same cluster by analyzing the characteristics of the target data. The characteristics of the target data may include the trend and distribution of the data values ​​included in the target data. For example, controller 120 may employ various algorithms, such as decision tree-based algorithms (e.g., XGBoost), K-Nearest Neighbor (K-NN) algorithms (e.g., K-means), and Support Vector Machine (SVM) algorithms, to cluster the target data. In an implementation, controller 120 may include a learning model to cluster feature data based on target data.

[0051] According to the implementation, controller 120 can classify the type of the battery dataset based on clustering results. The type of the battery dataset can be represented by a combination of characteristics that can identify the battery usage patterns and driving habits of users of vehicles equipped with batteries. For example, controller 120 can express the type of the battery dataset using a combination of multiple characteristics similar to the personality type represented by the Myers-Briggs Type Indicator (MBTI), which uses combinations of multiple characteristics to represent an individual's personality type. Controller 120 can represent user types not only by combinations of characteristics but also by levels represented by numbers, graphics, symbols, shapes, and various other methods.

[0052] By performing clustering based on the target data, the controller 120 can derive classification results in which the characteristics of each cluster are more clearly distinguished, thereby potentially improving the performance of type classification of the battery dataset using the enhanced clustering results.

[0053] Figure 2This is a diagram illustrating an example of feature data in the form of a contour image according to an embodiment disclosed in this document.

[0054] Reference Figure 2 The controller 120 can generate feature data representing the correlation between at least two factors in the form of a contour image, wherein each data value of the feature data can represent the joint probability density of the two factors, and the contour lines connect points with equal probability density values.

[0055] exist Figure 2 As an example, feature data representing the correlation between charging start time and charging start days in a week is shown. Here, charging start time can be represented in 24 units from 0 to 24 hours, and charging start days in a week can be represented in 7 units from Monday to Sunday. Therefore, the feature data can have (24... 7) Data size.

[0056] Figure 3a and Figure 3b This is a diagram illustrating an example of characteristic data based on the number of peak points according to the implementation disclosed in this document. Figure 3a The feature data with a peak point and the corresponding target data are shown. Figure 3b The feature data with two peak points and the corresponding target data are shown.

[0057] The controller 120 can determine the number of peak points for each feature data. Peak points can refer to the highest point in a contour image (e.g., the point with the highest probability density) or the highest contour line. For example, in... Figure 3a In the diagram, the highest contour line 310 can represent the peak point. Similarly, when in... Figure 3b When different contour patterns 320 and 330 exist in the feature data shown, the controller 120 can identify each highest contour line 340 and 350 in each contour pattern as a peak point.

[0058] The controller 120 can group feature data based on the number of peak points. For example, the controller 120 can... Figure 3a The feature data shown and Figure 3b The feature data shown is grouped into separate groups. That is, since the size of the target data varies depending on the number of peak points, the controller 120 can group the feature data based on the number of peak points to allow clustering using consistent criteria.

[0059] The controller 120 can extract target data based on the width, length, and centroid coordinates of each peak point of each feature data. For example, the controller 120 can extract data with (4 The target data for the number of peak points, such as Figure 3a and Figure 3b As shown in the image.

[0060] By extracting target data from feature data in this manner, controller 120 can reduce the size of the data to be clustered. For example, refer to Figure 2 , Figure 3a and Figure 3b It can be observed that as the controller 120 extracts the target data, the data size increases from (24... 7) Significantly reduced to (4) The number of peak points.

[0061] Figure 4 This is a flowchart illustrating the operation method of the data analysis apparatus according to the embodiments disclosed in this document. Figure 4 The embodiments shown are merely illustrative, and the order of operation according to various embodiments of the present invention may vary. Figure 4 The differences shown, and Figure 4 Some of the operations shown can be omitted, rearranged, or combined.

[0062] Reference Figure 4 The operation method of the battery analysis device may include acquiring multiple battery datasets in operation S100, generating feature data representing the correlation between at least two factors for each battery dataset in operation S200, extracting target data associated with at least one factor based on the characteristics of the feature data for each feature data in operation S300, clustering the feature data based on the target data in operation S400, and classifying the type of the battery dataset based on the clustering results in operation S500.

[0063] In operation S100, the information acquisition unit 110 can acquire multiple battery datasets. Each battery dataset may include time-series data of battery data. For example, the information acquisition unit 110 may receive battery data from the battery management system (BMS) installed in each vehicle's battery pack and from the on-board diagnostic (OBD) device equipped in the vehicle.

[0064] In operation S200, controller 120 can generate feature data for each battery dataset. Controller 120 can generate feature data representing the correlation between at least two factors. In one implementation, controller 120 can generate feature data based on the joint probability density function of at least two factors.

[0065] In operation S300, the controller 120 can extract target data for each feature data. The controller 120 can also extract target data associated with at least one parameter based on the characteristics of the feature data. In an embodiment, the feature data may include contour image data, and the at least one parameter may include the number of peak points in the contour image, the width, length, and coordinates of the center of each peak point.

[0066] In operation S400, controller 120 can cluster feature data based on target data. Controller 120 can extract smaller target data from the feature data and perform clustering based on the target data, thereby producing clustering results with more diverse characteristics.

[0067] In operation S500, controller 120 can classify the type of battery dataset based on clustering results. The type of battery dataset can be represented by a combination of characteristics that can identify the battery usage patterns and driving habits of users of vehicles equipped with batteries.

[0068] Figure 5 This is a flowchart illustrating the process of extracting target data by a data analysis device according to an embodiment disclosed in this document.

[0069] Reference Figure 5 Extracting target data in operation S300 may include: calculating the number of peak points of each feature data in operation S310, grouping the feature data based on the number of peak points in operation S320, and extracting target data based on the width, length and centroid coordinates of the peak points of each feature data.

[0070] In operation S310, controller 120 may calculate the number of peak points for each feature data. For example, controller 120 may identify the highest point in a contour image (e.g., the point with the highest probability density) as a peak point. In another example, a peak point may refer to a contour line representing the highest data value (i.e., the highest contour line).

[0071] In operation S320, controller 120 can group feature data based on the number of peak points. Since the size of the target data can vary depending on the number of peak points, controller 120 can group feature data with the same number of peak points to cluster the feature data using a consistent criterion.

[0072] In operation S330, controller 120 can extract target data based on the width, length, and centroid coordinates of each peak point of each feature data. For example, controller 120 can extract the width, length, x-coordinate of the centroid, and y-coordinate of the centroid of each peak point of each feature data. In this case, the target data of each feature data can have (4 The data size (number of peak points).

[0073] Figure 6 This is a diagram illustrating the hardware configuration of a computing system for performing an operation method of a data analysis apparatus according to an embodiment disclosed in this document.

[0074] Reference Figure 6 The computing system 1000 according to the embodiments disclosed in this document may include an MCU 1010, a memory 1020, an input and output I / F 1030, and a communication I / F 1040.

[0075] MCU 1010 can be a processor that executes various programs stored in memory 1020 (e.g., battery data acquisition program, feature data generation program, target data extraction program, and type classification program) and processes various information (including time-series data of the dataset) through these programs to perform reference... Figure 2 The data analysis device described includes the functions of the controller.

[0076] The memory 1020 can store various programs, such as battery data acquisition programs, feature data generation programs, target data extraction programs, and type classification programs. The memory 1020 can also store various information, including clustering results and type classification results.

[0077] If necessary, multiple memory units 1020 can be configured. Memory units 1020 can be volatile memory or non-volatile memory. Memory units 1020 that are volatile memory can be RAM, DRAM, SRAM, etc. Memory units 1020 that are non-volatile memory can be ROM, PROM, EAROM, EPROM, EEPROM, flash memory, etc. Memory units 1020 are not limited to the listed examples, nor are they limited to these examples.

[0078] The Input and Output I / F 1030 is an interface for connecting input devices (not shown), such as a keyboard, mouse, or touch panel, output devices (not shown), such as a display, and the MCU 1010 to send and receive data.

[0079] The communication I / F 1040 can be a component capable of transmitting various types of data to a server, including various devices that support wired or wireless communication. For example, a data analysis device can send various types of information to and receive various types of information from a separate external server via the communication interface 1040.

[0080] In this way, a computer program according to the embodiments disclosed in this document can be recorded in memory 1020 and processed by MCU 1010 to be implemented as an execution reference. Figure 1 The module that describes the functionality.

[0081] Although all components are described as operating in combination or as a single component, the implementations disclosed in this document are not limited to these examples. That is, within the scope of the implementations disclosed in this document, all components may also be selectively combined in one or more ways to function.

[0082] Furthermore, unless otherwise expressly stated, terms such as “comprising,” “including,” or “having” as used above should be interpreted as implying the inclusion of specified components rather than excluding other components, and may include additional components. Unless otherwise defined, all terms (including technical or scientific terms as used herein) have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments disclosed in this invention pertain. It should also be understood that terms such as those defined in common dictionaries should be interpreted as having the meaning consistent with their meaning in the context of the relevant field, and should not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0083] The above description is merely an illustrative example of the technical concepts disclosed in this document, and those skilled in the art will be able to make various modifications and changes to the embodiments disclosed in this document without departing from the subject matter of the embodiments disclosed in this document. Therefore, the embodiments disclosed in this document are not intended to limit the technical concepts disclosed in this document, but are for illustrative purposes, and the scope of the technical concepts disclosed in this document is not limited by these embodiments. The scope of protection of the technical concepts disclosed in this document should be interpreted in accordance with the claims set forth herein, and all equivalent technical concepts should be considered to be included within the scope of protection of this document.

Claims

1. A data analysis device, the data analysis device comprising: An information acquisition unit, configured to acquire multiple battery datasets; as well as A controller, configured to generate feature data for each battery dataset representing the correlation between at least two factors used to analyze the type of the battery dataset, and For each feature data, target data associated with at least one parameter is extracted based on the characteristics of the feature data, and the feature data is clustered based on the target data.

2. The data analysis device according to claim 1, wherein, The feature data includes contour image data.

3. The data analysis device according to claim 2, wherein, The at least one parameter includes at least one of the following: the number of peak points in the contour image data, the width and length of each peak point, and the centroid coordinates.

4. The data analysis device according to claim 3, wherein, The controller calculates the number of peak points for each feature data and groups the feature data based on the number of peak points.

5. The data analysis apparatus according to claim 4, wherein, The controller extracts the target data based on the width, length, and centroid coordinates of the peak points of each feature data.

6. The data analysis apparatus according to claim 4, wherein, The controller clusters the feature data belonging to each group based on the target data.

7. The data analysis apparatus according to claim 1, wherein, The controller classifies the battery dataset based on clustering results.

8. The data analysis apparatus according to claim 1, wherein, The controller generates the feature data based on the joint probability density function (joint PDF) of the at least two factors.

9. A method for operating a data analysis device, wherein, The operation method of the data analysis device includes the following steps: Obtain multiple battery datasets; For each battery dataset, feature data is generated representing the correlation between at least two factors used to analyze the type of the battery dataset; For each feature data, target data associated with at least one parameter is extracted based on the characteristics of the feature data; and Cluster the feature data based on the target data.

10. The method of operating the data analysis device according to claim 9, wherein, The feature data includes contour image data, and the at least one parameter includes at least one of the following: the number of peak points in the contour image data, the width of each peak point, the length of each peak point, and the centroid coordinates.

11. The method of operating the data analysis device according to claim 10, wherein, The steps for calculating the target data include the following: Calculate the number of peak points for each feature data; The feature data is grouped based on the number of peak points; and The target data is extracted based on the width, length, and centroid coordinates of the peak points of each feature data.

12. The method of operating the data analysis device according to claim 11, wherein, The steps of clustering the feature data based on the target data include: For each group, the feature data belonging to each group are clustered based on the target data.

13. The operation method of the data analysis device according to claim 9, wherein the operation method of the data analysis device further includes classifying the type of the battery dataset based on the clustering results.

Images