Information processing method

Abstract

To provide an information processing method that can find out a setting in which extracted data capturing a feature of the entire original data can be obtained.SOLUTION: An information processing method includes: a first step (S120) of calculating a relative frequency distribution in original data for each feature amount; a second step (S125) of setting a plurality of time windows to cut out data corresponding to part of a period from the original data; a third step (S130) of cutting out the data from the original data; a fourth step (S140) of calculating a relative frequency distribution in extracted data; and a fifth step (S145) of calculating an error between the relative frequency distribution in the original data and the relative frequency distribution in the extracted data. An information processor repeatedly executes trials from the second step to the fifth step while changing the setting of the plurality of time windows after executing the first step, and selects and outputs settings of the plurality of time windows in which the error is less than a threshold value.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to an information processing method. [Background technology]

[0002] Patent Document 1 discloses an information processing device that reduces the size of original data for analysis by compressing the data. The original data for analysis is data collected over a predetermined period of time using a sensor mounted on a vehicle.

[0003] The information processing device disclosed in Patent Document 1 compresses data by extracting, from original data, data acquired when a certain vehicle speed is reached and data acquired at a vehicle speed inflection point. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2008-108247 A Summary of the Invention [Problem to be solved by the invention]

[0005] The above information processing device extracts data by focusing only on the vehicle speed. Therefore, the above information processing device cannot extract data according to the characteristics of the data other than the vehicle speed. There is a demand for an information processing method that can obtain extracted data that captures the characteristics of the entire original data. [Means for solving the problem]

[0006] The means for solving the above problems and their effects will be described below. An information processing method for solving the above problem is an information processing method that reduces the amount of data used for analysis by extracting a portion of data from original data collected over a predetermined period of time using multiple sensors installed in a vehicle.

[0007] This information processing method includes a first step in which an information processing device calculates a relative frequency distribution in the original data for each of a plurality of feature amounts contained in the original data.

[0008] This information processing method includes a second step in which the information processing device sets a plurality of time windows that extract data for a portion of the original data so that the sum of the periods of all of the time windows is shorter than the default period.

[0009] The information processing method includes a third step in which the information processing device extracts data from the original data using the plurality of time windows. This information processing method includes a fourth step in which the information processing device calculates, for each of the feature amounts, the relative frequency distribution in extracted data obtained by combining all of the data extracted using the plurality of time windows.

[0010] This information processing method includes a fifth step in which the information processing device calculates an error between the relative frequency distribution in the original data and the relative frequency distribution in the extracted data.

[0011] In this information processing method, after executing the first step, the information processing device repeatedly executes trials from the second step to the fifth step by changing the settings of the multiple time windows, and selects and outputs settings of the multiple time windows that make the error less than a threshold value.

[0012] In one aspect of the information processing method, the information processing device further includes a step of executing clustering, which is machine learning, for classifying data of each section obtained by dividing the original data into a certain period of time into a predetermined number of clusters. In the second step, the information processing device sets the multiple time windows so that the ratio of each cluster in the extracted data is equal to the ratio of each cluster in the entire original data. Effect of the Invention

[0013] According to the above information processing method, it is possible to find, using an information processing device, settings that can produce extracted data that captures the overall characteristics of the original data. [Brief description of the drawings]

[0014] [Figure 1] FIG. 1 is a schematic diagram showing the relationship between a data center that executes an information processing method, a vehicle, and an information processing terminal. [Diagram 2] FIG. 2 is a graph showing the original data, where (a) shows the change in vehicle speed, (b) shows the change in inclination angle, and (c) shows the change in acceleration. [Diagram 3] FIG. 3 is a flowchart showing the flow of processing executed by the processing device in the data center. [Figure 4] FIG. 4 is a graph showing an example of clustering of original data using two feature amounts. [Diagram 5] FIG. 5 is a graph showing an example of a relative frequency distribution of vehicle speed in the original data. [Figure 6] FIG. 6 is a graph showing an example of a relative frequency distribution for the tilt angle in the original data. [Figure 7] FIG. 7 is a flowchart showing a part of the flow of processing executed by a processing device that executes an information processing method according to the modified example. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] Hereinafter, an embodiment of an information processing method and an information processing device that executes the information processing method will be described with reference to FIGS. 1 to 6. FIG. <Configuration of information processing system> Fig. 1 shows a configuration of an information processing system including a data center 500 equipped with an information processing device according to an embodiment. As shown in Fig. 1, the data center 500 communicates with a plurality of vehicles 10 via a communication network 400. The data center 500 also communicates with an information processing terminal 600 via the communication network 400.

[0016] <Configuration of Data Center 500> As shown in FIG. 1, the data center 500 includes a processing device 510 as an information processing device. The data center 500 includes a storage device 520 and a communication device 530. The processing device 510 includes a CPU that executes processing according to a program, and a ROM in which the program is stored. The storage device 520 stores a large amount of data. The communication device 530 is implemented as hardware such as a network adapter, various communication software, or a combination of these. The communication device 530 realizes wired or wireless communication via the communication network 400.

[0017] The data center 500 may be configured using a plurality of computers. For example, the data center 500 may be configured by a plurality of server devices. <Vehicle 10 Configuration> Each of the multiple vehicles 10 is equipped with a communication device 80. These communication devices 80 are implemented as hardware such as a network adapter, various communication software, or a combination of these. These communication devices 80 are configured to realize wired or wireless communication via a communication network 400. Each vehicle 10 is equipped with various sensors that collect information from each part of the vehicle 10.

[0018] In each vehicle 10, driving data is collected from these various sensors. The driving data is transmitted from each vehicle 10 to the data center 500 by the communication device 80. For example, driving data including the driving distance, location information, and vehicle speed of each vehicle 10 is transmitted to the data center 500. Identification information for identifying each vehicle 10 is also transmitted to the data center 500 together with the driving data.

[0019] The data center 500 stores the received identification information and the travel data in the storage device 520. In this manner, the travel data of the multiple vehicles 10 is accumulated in the storage device 520 of the data center 500.

[0020] <Configuration of information processing terminal 600> The information processing terminal 600 includes a processing device 610, a storage device 620, and a communication device 630. The processing device 610 includes a CPU that executes processing according to a program, and a ROM in which the program is stored. The storage device 620 stores a large amount of data. The communication device 630 is implemented as hardware such as a network adapter, various communication software, or a combination of these. The communication device 630 realizes wired or wireless communication via the communication network 400.

[0021] The information processing terminal 600 is, for example, a personal computer. <Analysis of vehicle 10's driving data> The information processing terminal 600 is used to analyze the driving data. When analyzing the driving data, a portion of the driving data is downloaded to the storage device 620 of the information processing terminal 600 from the vast amount of driving data stored in the storage device 520 of the data center 500. The driving data to be downloaded is selected according to the purpose of the analysis. The processing device 610 performs analysis using the downloaded driving data.

[0022] For example, the processing device 610 calculates a load on a specific part of the vehicle 10 based on the driving data. The processing device 610 estimates damage accumulated in the part based on the calculated load. For example, the processing device 610 performs an analysis to estimate durability required for a specific part of the vehicle 10 based on the driving data of a plurality of vehicles 10.

[0023] To perform such an analysis, the processing unit 610 analyzes a large amount of driving data collected over a long period of time. Since the processing unit 610 needs to perform a huge amount of calculations, the analysis takes a long time.

[0024] Therefore, it is conceivable to extract extracted data that captures the characteristics of the entire original data from a large amount of driving data, which is the original data. If such extracted data can be extracted, the processing device 610 can perform analysis in a shorter time by using the extracted data. For example, when estimating damage to a part after 100,000 hours of driving, the processing device 610 estimates the damage using 20,000 hours of extracted data extracted from 100,000 hours of original data. Then, the processing device 610 calculates an estimated value of damage to the part after 100,000 hours of driving by multiplying the estimated value calculated from the 20,000 hours of extracted data by five.

[0025] Fig. 2 shows an example of original data. The original data shown in Fig. 2 is driving data for 100,000 hours for one vehicle 10. The original data shown in Fig. 2 includes vehicle speed, inclination angle, and acceleration as feature amounts.

[0026] FIG. 2(a) shows the progress of vehicle speed over 100,000 hours. FIG. 2(b) shows the progress of inclination angle of vehicle 10 over 100,000 hours. The inclination angle is a positive value when going uphill. The inclination angle is a negative value when going downhill. FIG. 2(c) shows the acceleration of vehicle 10. The acceleration is a positive value when vehicle 10 is accelerating. The acceleration is a negative value when vehicle 10 is decelerating.

[0027] The vehicle speed, the inclination angle, and the acceleration are correlated with the load on the powertrain of the vehicle 10. The processing device 610 of the information processing terminal 600 estimates the load on the components that constitute the powertrain from the driving data including the vehicle speed, the inclination angle, and the acceleration as feature quantities.

[0028] When transmitting the data for analysis to the information processing terminal 600, the data center 500 also transmits information for extracting the extracted data from the original data together with the original data.

[0029] The extracted data is created by extracting data from the original data using a plurality of time windows. In FIG. 2, five time windows, a first time window W_1, a second time window W_2, a third time window W_3, a fourth time window W_4, and a fifth time window W_5, are shown by dashed lines as examples of the plurality of time windows. The start and end of each time window are set so that the time windows do not overlap. In this example, 20,000 hours of driving data are extracted as the extracted data. Therefore, the start and end of each time window are set so that the total length of the periods of all the time windows W_1 to W_5 is 20,000 hours.

[0030] The data center 500 searches for settings of the start and end of each time window that indicates an extraction pattern for extracting extracted data that captures the characteristics of the entire original data. The data center 500 transmits information on the extraction pattern for extracting the extracted data described above to the information processing terminal 600 together with the original data. In other words, the information on the extraction pattern for extracting the extracted data that is transmitted together with the original data is information on the settings of each time window found by the search.

[0031] The information processing terminal 600 receives information on an extraction pattern for extracting extracted data from the original data together with the original data, and sets the extraction pattern based on the received information. The information processing terminal 600 then extracts the extracted data from the original data using the set extraction pattern. The information processing terminal 600 performs analysis using the extracted data.

[0032] <Searching for Extraction Patterns> 3 is a flowchart showing a series of processes related to the extraction pattern search process. This series of processes is executed by the processing device 510 of the data center 500.

[0033] 3, the processing device 510 acquires original data in the process of step S100. The original data is a portion of the traveling data selected according to the purpose of the analysis from the huge amount of traveling data stored in the storage device 520 of the data center 500. For example, the original data for calculating the load on a specific part of one vehicle 10 is the traveling data of the target vehicle 10 over a predetermined period selected from the huge amount of traveling data of multiple vehicles 10.

[0034] In the process of step S110, the processing device 510 assigns labels to the original data by clustering. Specifically, the processing device 510 divides the original data into certain periods. The length of the periods into which the original data is divided is, for example, several minutes. Then, the processing device 510 executes clustering, which is machine learning, for classifying the data of each period into a predetermined number of clusters. For example, the k-means method is used as the clustering algorithm. The k-means method is a clustering algorithm that classifies data into a pre-specified number of clusters. The clustering algorithm is not limited to the k-means method.

[0035] The original data includes driving data collected under different environments, such as driving data when driving in urban areas, driving data when driving in suburban areas, and driving data when driving on expressways. By performing clustering, the driving data included in the original data can be classified into clusters of driving data with similar characteristics. The number of clusters to be classified is set arbitrarily depending on the content of the analysis.

[0036] FIG. 4 is a graph showing an example of clustering original data into four clusters by the k-means method using two feature amounts contained in the original data as explanatory variables. In FIG. 4, each data section into which the original data is divided is shown by one point. When performing clustering, the processing device 510 uses a representative value of the explanatory variable in the data of each section. For example, the processing device 510 sets the average value of the feature amounts in the data of each section as the representative value. The processing device 510 may use a moving average value of the feature amounts in a plurality of sections that are consecutive in time series as the representative value.

[0037] In FIG. 4, these points are shown in a two-dimensional space with the first feature FV_a and the second feature FV_b as coordinate axes. FIG. 4 shows an example of original data clustered into four clusters: a first cluster M_1, a second cluster M_2, a third cluster M_3, and a fourth cluster M_4. In FIG. 4, the boundaries of the four clusters are shown with solid lines. In FIG. 4, the centers of gravity of each cluster are shown with hollow triangles. The center of gravity cgM_1 is the center of gravity of the first cluster M_1. The center of gravity cgM_2 is the center of gravity of the second cluster M_2. The center of gravity cgM_3 is the center of gravity of the third cluster M_3. The center of gravity cgM_4 is the center of gravity of the fourth cluster M_4.

[0038] 4 shows an example in which there are two explanatory variables, but the number of explanatory variables is not limited to two. For example, as in the example shown in FIG. 2, when the original data includes three feature amounts, namely, vehicle speed, tilt angle, and acceleration, the processing device 510 may perform clustering using these three feature amounts as explanatory variables. In this case, the processing device 510 performs clustering on the original data in a three-dimensional coordinate space.

[0039] The processing device 510 assigns a label indicating the result of the clustering to the original data. Specifically, the processing device 510 assigns a label that identifies the cluster into which the data is classified to each piece of data that is represented by a point in the coordinate space. In this way, the processing device 510 creates original data with the label.

[0040] Next, in the process of step S120, the processing device 510 calculates the relative frequency distribution of the original data. As described above, the original data includes a plurality of feature amounts. The processing device 510 calculates the relative frequency distribution in the original data for each feature amount.

[0041] A frequency distribution classifies data into multiple classes and shows the distribution of the frequency, which is the number of data in each class. Relative frequency indicates what percentage the frequency of that class makes up of the total frequency.

[0042] Fig. 5 shows the relative frequency distribution of the vehicle speed in the original data shown in Fig. 2. In this relative frequency distribution, the vehicle speed in the original data is divided into m classes from 1 to m, and the relative frequency distribution is shown.

[0043] Fig. 6 shows the relative frequency distribution for the tilt angles in the original data shown in Fig. 2. In this relative frequency distribution, the tilt angles in the original data are divided into m classes from 1 to m, and the relative frequency distribution is shown.

[0044] In the process of step S120, the processor 510 calculates such a relative frequency distribution for each feature amount included in the original data. The number of classes in the relative frequency distribution for each feature amount is the same.

[0045] For example, as shown in the example of FIG. 2, when the original data includes three feature amounts, namely, vehicle speed, tilt angle, and acceleration, the processing device 510 calculates the relative frequency distribution of each of these three feature amounts.

[0046] Next, in the process of step S125, the processor 510 sets a plurality of time windows to extract extracted data from the original data. In Fig. 2, five time windows W_1 to W_5, namely, a first time window W_1, a second time window W_2, a third time window W_3, a fourth time window W_4, and a fifth time window W_5, are shown as an example of a plurality of time windows. In the example shown in Fig. 2, the durations of the time windows are all equal. As shown in Fig. 2, data extracted by each extraction window is data of each feature amount in the same period.

[0047] In the process of step S125, the processing device 510 randomly sets a plurality of time windows so that the sum of the periods of all the time windows is shorter than a predetermined period, which is the period of the entire original data. As will be described later, the processing device 510 creates extracted data by combining all the data cut out by the plurality of time windows set here. The sum of the periods of all the time windows is a value that determines the capacity of the extracted data. Therefore, the sum of the periods of all the time windows is set in advance.

[0048] For example, the processing device 510 randomly sets the number of time windows, the start time of each time window, and the end time of each time window every time the processing of step S125 is executed. At this time, the processing device 510 sets each time window so that the time windows do not overlap. The processing device 510 thus randomly sets a plurality of time windows so that the sum of the periods of all the time windows is a period set in advance. In the processing of step S125, the processing device 510 may set a plurality of time windows by fixing the period of each time window to a constant value as shown in FIG. 2. In the processing of step S125, the processing device 510 may set a plurality of time windows by fixing the number of the plurality of time windows to a constant value.

[0049] In addition to the above requirements, when setting multiple time windows through step S125, the processing device 510 sets the multiple time windows so that the proportion of each cluster in the extracted data is equal to the proportion of each cluster in the entire original data.

[0050] In this way, a cut-out pattern for cutting out data from the original data is determined by setting a plurality of time windows through the process of step S125. After determining the cut-out pattern in this way, the processing device 510 advances the process to step S130.

[0051] In the process of step S130, the processing device 510 extracts data from the original data using the determined extraction pattern. That is, in the process of step S130, the processing device 510 extracts data from the original data using the multiple time windows that have been set. Then, the processing device 510 combines all the data extracted using the multiple time windows to create extracted data.

[0052] In the process of the next step S140, the processing device 510 calculates the relative frequency distribution of the extracted data. The processing device 510 calculates the relative frequency distribution of the extracted data in the same manner as the method for calculating the relative frequency distribution in step S120. That is, in the process of step S140, the processing device 510 calculates the relative frequency distribution in the extracted data for each feature amount. At this time, the processing device 510 sets the number of classes in the relative frequency distribution of each feature amount to be the same as the relative frequency distribution in step S120.

[0053] For example, as in the example shown in FIG. 2, when the original data includes three feature amounts, namely, vehicle speed, tilt angle, and acceleration, the processing device 510 also calculates the relative frequency distribution of each of these three feature amounts in step S140.

[0054] Next, in the process of step S145, the processing device 510 calculates the error between the relative frequency distribution in the original data and the relative frequency distribution in the extracted data. For example, the processing device 510 calculates the mean absolute error (MAE). The mean absolute error (MAE) is expressed by the following formula.

[0055]

number

[0056] In the above formula, "n" is the number of features. "m" is the number of classes in the relative frequency distribution. "Y" is the frequency of the corresponding feature in the corresponding class in the original data. "y" is the frequency of the corresponding feature in the corresponding class in the extracted data.

[0057] As shown in the above formula, the processing device 510 calculates the error as the sum of the errors in the frequencies in each class for each feature between the relative frequency distribution in the entire original data and the relative frequency distribution in the extracted data.

[0058] After calculating the error, the processing device 510 advances the process to step S150. In the process of step S150, the processing device 510 judges whether the calculated error is equal to or smaller than a threshold value. The threshold value is a value for judging whether extracted data having a relative frequency distribution close to the relative frequency distribution in the original data has been extracted by the set extraction pattern. The magnitude of this threshold value is set in advance so that it can be judged that extracted data having a relative frequency distribution close to the relative frequency distribution in the original data has been extracted based on the error being equal to or smaller than the threshold value.

[0059] In the process of step S150, when it is determined that the error is equal to or smaller than the threshold value (step S150: YES), the processing device 510 advances the process to step S160. In the process of step S160, the processing device 510 records the cut-out pattern. Specifically, the processing device 510 stores data on the start and end of each time window in the cut-out pattern in the storage device 520 as information specifying the cut-out pattern. After recording the cut-out pattern in this manner, the processing device 510 advances the process to step S170.

[0060] On the other hand, when it is determined in the process of step S150 that the error is greater than the threshold value (step S150: NO), processing device 510 does not execute the process of step S160 and proceeds to step S170. That is, processing device 510 proceeds to the process of step S170 without recording the extraction pattern.

[0061] In the process of step S170, the processing device 510 judges whether or not the number of trials is equal to or greater than a preset number. The number of trials is the number of times trials consisting of the processes of steps S125 to S145 have been performed. These trials are trials to search for an extraction pattern for extracting extracted data. The preset number is a threshold for judging whether a sufficient number of trials to search for an extraction pattern for extracting extracted data have been performed. This preset number is set in advance so that it can be judged that an extraction pattern has been sufficiently searched for based on the fact that the number of trials is equal to or greater than the preset number.

[0062] In the process of step S170, when it is determined that the number of trials is less than the predetermined number (step S170: NO), the processing device 510 returns the process to step S125 and performs a trial again. That is, the processing device 510 performs a trial of setting a plurality of new time windows and extracting extracted data.

[0063] On the other hand, if it is determined in the process of step S170 that the number of attempts is equal to or greater than the predetermined number of attempts (step S170: YES), processor 510 advances the process to step S180.

[0064] In this manner, the processing device 510 repeats the trial until the number of trials reaches the preset number. The extraction pattern in which the error is thus equal to or smaller than the threshold is stored in the storage device 520. When the number of trials reaches the preset number, the processing device 510 advances the process to step S180.

[0065] In the process of step S180, processing device 510 outputs the cut-out pattern stored in storage device 520. Specifically, processing device 510 transmits the setting of the time window in which the error is equal to or smaller than the threshold value, which is recorded in storage device 520, to information processing terminal 600 as a candidate for the cut-out pattern. When the cut-out pattern is output in the process of step S180, processing device 510 ends this series of processes.

[0066] Thus, through this series of processes, the processing device 510 selects and outputs all time window settings that have an error equal to or less than the threshold value from among the trialed time window settings. <Action of this embodiment> The processing device 510 of the data center 500 executes the information processing method for searching for settings for reducing the amount of data used in the analysis, as described above.

[0067] The information processing method of this embodiment includes a first step (step S120) in which the processing device 510 calculates a relative frequency distribution in the original data for each of a plurality of feature amounts included in the original data. The information processing method includes a second step (step S125) in which the processing device 510 sets a plurality of time windows so that the total period of all the time windows is shorter than a default period. The information processing method includes a third step (step S130) in which the processing device 510 extracts data from the original data using a plurality of time windows. The information processing method includes a fourth step (step S140) in which the processing device 510 calculates, for each of the feature amounts, a relative frequency distribution in extracted data obtained by combining all the data extracted using the plurality of time windows. The information processing method includes a fifth step (step S145) in which the processing device 510 calculates an error between the relative frequency distribution in the original data and the relative frequency distribution in the extracted data.

[0068] In this information processing method, the processing device 510 executes the first step, and then repeatedly executes trials from the second step to the fifth step by changing the settings of a plurality of time windows. Then, the processing device 510 selects and outputs the settings of a plurality of time windows that make the error less than a threshold value.

[0069] In this information processing method, the processing device 510 calculates the error between the relative frequency distribution of each feature in the entire original data and the relative frequency distribution of each feature in the extracted data. The processing device 510 repeats this error calculation while changing the settings of multiple time windows for extracting data from the original data. The processing device 510 then selects and outputs the settings that make the error less than a threshold.

[0070] According to this information processing method, the time window setting can be searched for by the processing device 510. The setting output from the processing device 510 by this information processing method is a setting that can extract extracted data having a relative frequency distribution of each feature amount close to that of the entire original data.

[0071] <Effects of this embodiment> (1) According to the above information processing method, it is possible to find settings that can produce extracted data that captures the overall characteristics of the original data.

[0072] (2) In the above example, the original data includes vehicle speed, tilt angle, and acceleration as feature quantities. In this case, the processing device 510 calculates a relative frequency distribution for vehicle speed in the original data in the above first step. In this case, the processing device 510 calculates a relative frequency distribution for tilt angle in the original data in the above first step. In this case, the processing device 510 calculates a relative frequency distribution for acceleration in the original data in the above first step.

[0073] In this case, in the fourth step, the processing device 510 calculates a relative frequency distribution for the vehicle speed in the extracted data. In this case, in the fourth step, the processing device 510 calculates a relative frequency distribution for the tilt angle in the extracted data. In this case, in the fourth step, the processing device 510 calculates a relative frequency distribution for the acceleration in the extracted data.

[0074] The vehicle speed, the tilt angle, and the acceleration are correlated with the load on the power train of the vehicle 10. By obtaining extracted data including the vehicle speed, the tilt angle, and the acceleration, the load on the power train of the vehicle 10 can be estimated with a small amount of data.

[0075] This allows the processing unit 510 to find settings that can provide extracted data for use in estimating the load on the powertrain of the vehicle 10 . (3) The above information processing method includes a step (step S110) in which the processing device 510 executes clustering, which is machine learning, to classify data of each section obtained by dividing the original data into a certain period of time into a predetermined number of clusters. In the second step (step S125), the processing device 510 sets a plurality of time windows so that the ratio of each cluster in the extracted data is equal to the ratio of each cluster in the entire original data.

[0076] A plurality of sections classified into the same cluster are sections with similar characteristics. The settings output from the processing device 510 by the above information processing method are settings that can extract extracted data in which the ratio of each cluster to the entire original data is equal and the relative frequency distribution of each feature is close.

[0077] Therefore, according to the above information processing method, it is possible to find settings that can obtain extracted data that is closer to the characteristics of the entire original data. (4) In the above information processing method, the processing device 510 selects and outputs all time window settings for which the error is equal to or smaller than a threshold value from among the trialed time window settings.

[0078] In the above information processing method, all extraction patterns with an error equal to or less than a threshold are output, so that the multiple output extraction patterns can be compared and one extraction pattern to be finally adopted can be selected from among them.

[0079] <Example of change> This embodiment can be modified as follows: This embodiment and the following modifications can be combined with each other to the extent that there is no technical contradiction.

[0080] In the above example, the processing device 510 sets a plurality of time windows so that the ratio of each cluster in the original data is equal to the ratio of each cluster in the extracted data. The processing device 510 may set a plurality of time windows without imposing such a constraint. In this case, the process of step S110 for performing clustering may be omitted.

[0081] In the above information processing method, in the fifth step, the processing device 510 calculates the interval error, which is the difference in relative frequency for each class between the relative frequency distribution of the original data and the relative frequency distribution of the extracted data, and simply calculates the sum of the interval errors. Alternatively, each calculated interval error may be adjusted by multiplying it by a weight, and then the sum of the interval errors may be calculated as the error.

[0082] 7, the processor 510 calculates the relative frequency distribution of the extracted data in step S140, and then proceeds to step S142. In step S142, the processor 510 calculates an interval error, which is the difference in relative frequency for each class between the relative frequency distribution of the original data and the relative frequency distribution of the extracted data.

[0083] Next, in the process of step S144, the processor 510 adjusts the interval error for each class by multiplying the calculated interval error by a weight for each class. Then, in the process of step S146, the processor 510 calculates the sum of the adjusted interval errors as the error. The processor 510 executes the processes from step S150 onward using the error thus calculated.

[0084] Depending on the feature, there may be a large bias in the frequency of occurrence for each class. According to the information processing method in which each interval error is adjusted by multiplying it by a weight as described above, the weight of the interval error is adjusted to suppress the influence of the bias in the frequency of occurrence and calculate the error.

[0085] The method of determining the setting of the time window in the cut-out pattern does not have to be random. The setting of the time window in the cut-out pattern may be changed according to a preset rule and repeated trials may be performed.

[0086] The error calculated in the process of step S145 is not limited to the mean absolute error MAE. For example, the processing device 510 may calculate the mean square error as the error. The processing device 510 may calculate the root mean square error as the error.

[0087] In the information processing method of the above embodiment, in the process of step S180, all the extraction patterns stored in the storage device 520 are output. In contrast, the information processing method does not have to output all the extraction patterns. For example, the information processing method may be a method of selecting and outputting the extraction pattern with the smallest error from among the extraction patterns stored in the storage device 520.

[0088] In the above embodiment, an example has been shown in which the information processing device is embodied as the processing device 510 of the data center 500. Alternatively, the information processing device may be embodied as the processing device 610 of the information processing terminal 600. In this case, the processing device 610 downloads original data from the data center 500 and executes the series of processes shown in Figs. 3 and 7. [Explanation of symbols]

[0089] 10...vehicle, 80...communication device, 400...communication network, 500...data center, 510...processing device, 520...storage device, 530...communication device, 600...information processing terminal, 610...processing device, 620...storage device, 630...communication device

Claims

1. An information processing method for reducing the amount of data used for analysis by extracting a portion of data from original data collected over a predetermined period of time using a plurality of sensors mounted on a vehicle, a first step of an information processing device calculating a relative frequency distribution in the original data for each of a plurality of feature amounts included in the original data; a second step of setting a plurality of time windows by the information processing device, the time windows being cut out to extract data of a portion of the original data such that a total period of all the time windows is shorter than the predetermined period; a third step of the information processing device extracting data from the original data using the plurality of time windows; a fourth step in which the information processing device calculates, for each of the feature quantities, the relative frequency distribution in extracted data obtained by combining all of the data extracted using the plurality of time windows; a fifth step of calculating an error between the relative frequency distribution in the original data and the relative frequency distribution in the extracted data by the information processing device; the information processing device executes the first step, and then repeatedly executes trials from the second step to the fifth step by changing settings of the multiple time windows, and selects and outputs settings of the multiple time windows that make the error less than a threshold value.

2. The plurality of feature amounts include a vehicle speed, a tilt angle, and an acceleration. the information processing device calculates, in the first step, a relative frequency distribution for a vehicle speed in the original data, a relative frequency distribution for a tilt angle in the original data, and a relative frequency distribution for an acceleration in the original data; In the fourth step, the information processing device calculates a relative frequency distribution for a vehicle speed in the extracted data, a relative frequency distribution for a tilt angle in the extracted data, and a relative frequency distribution for an acceleration in the extracted data. The information processing method according to claim 1 .

3. The information processing device further includes a step of executing clustering, which is machine learning, in which data of each section obtained by dividing the original data into a certain period is classified into a predetermined number of clusters; In the second step, the information processing device sets the plurality of time windows so that a ratio of each cluster in the extracted data is equal to a ratio of each cluster in the entire original data. The information processing method according to claim 1 .

4. The information processing device selects and outputs all of the time window settings in which the error is equal to or smaller than the threshold value from among the trialed time window settings. The information processing method according to claim 1 .

5. In the fifth step, the information processing device calculates interval errors, which are differences in relative frequency for each class between the relative frequency distribution of the original data and the relative frequency distribution of the extracted data, and adjusts each calculated interval error by multiplying it by a weight, and then calculates a sum of the interval errors as the error. The information processing method according to claim 1 .

Images