JP7687388B2 - DATA ANALYSIS APPARATUS, DATA ANALYSIS METHOD, AND PROGRAM - Google Patents

Description

The present invention relates to a data analysis device, a data analysis method, and a recording medium, and in particular to a data analysis device, a data analysis method, and a recording medium for analyzing time series data.

Technologies for monitoring the operating status of equipment and parts in railroad cars, automobile engine rooms, factories, etc. are being researched. For example, the related technology described in Patent Document 1 uses a learning model generated by machine learning to diagnose the operating status of a machine in real time based on physical quantities detected by a sensor.

In one example of another related technique, non-negative matrix factorization (NMF) is used to analyze time-series data. Specifically, the related technique converts the time-series data into an amplitude spectrogram, and then decomposes the spectrogram into a basis matrix and an activation matrix. The activation matrix is then used as an acoustic feature to identify allophones contained in the time-series data.

JP 2020-204937 A

NMF approximates a non-negative matrix, which represents a spectrogram, by the product of lower-dimensional non-negative matrices. Therefore, when the period of peaks in time series data is unstable or in a noisy environment, the accuracy of identifying allophones in related techniques using NMF decreases.

The present invention has been made in consideration of the above problems, and its purpose is to accurately identify abnormal sounds from time series data of various characteristics.

A data analysis device according to one aspect of the present invention includes a determination means for determining the characteristics of time series data, a selection means for selecting a method for analyzing the time series data based on the characteristics of the time series data, and an identification means for identifying abnormal sounds contained in the time series data by analyzing the time series data using the selected method.

In one aspect of the present invention, a data analysis method determines the characteristics of time series data, selects a method for analyzing the time series data based on the characteristics of the time series data, and analyzes the time series data using the selected method to identify abnormal sounds contained in the time series data.

A recording medium according to one aspect of the present invention stores a program for causing a computer to execute the following steps: determine the characteristics of time series data; select a method for analyzing the time series data based on the characteristics of the time series data; and identify abnormal sounds contained in the time series data by analyzing the time series data using the selected method.

According to one aspect of the present invention, abnormal sounds can be accurately identified from time series data of various characteristics.

1 is a block diagram showing a configuration of a data analysis apparatus according to a first embodiment. 4 is a flowchart illustrating an operation of the data analysis apparatus according to the first embodiment. FIG. 11 is a block diagram showing a configuration of a data analysis apparatus according to a second embodiment. FIG. 11 is a diagram showing an example of time-series data to be analyzed by the data analysis apparatus according to the second embodiment. 10 is a flowchart illustrating an operation of a data analysis device according to a second embodiment. FIG. 11 is a block diagram showing a configuration of a data analysis apparatus according to a third embodiment. FIG. 13 is a diagram showing an example of a spectrogram converted from time series data. 1 is a graph showing an example of a frequency spectrum transformed from a segment of time series data of a certain duration. 13 is a flowchart illustrating an operation of a data analysis device according to a third embodiment. 1 is an example of a graph showing a distribution of scores used to determine a threshold value for a peak intensity of a frequency spectrum. FIG. 2 is a diagram illustrating an example of a hardware configuration of a data analysis apparatus according to any one of the first to third embodiments.

Several forms for implementing the present invention are described below.

[Embodiment 1]
A first embodiment will be described with reference to FIGS.

(Abnormal sound identification device 10)
Fig. 1 is a block diagram showing the configuration of an abnormal sound identification device 10 according to embodiment 1. As shown in Fig. 1, the abnormal sound identification device 10 includes a determination unit 11, a selection unit 12, a recognition unit 13, and a provision unit 14.

The determination unit 11 determines the nature of the time series data. The determination unit 11 is an example of a determination means.

In the first example, the determination unit 11 tracks the peaks of the time series data. Here, the determination unit 11 can use a well-known peak tracking technique. The determination unit 11 measures the time from the first peak to the next peak of the time series data. Then, the determination unit 11 calculates the time width (called a period) between adjacent peaks of the time series data. Then, the determination unit 11 calculates the fluctuation of the period of the peaks in the time series data. For example, the determination unit 11 calculates the variance or standard deviation of the period of the peaks in the time series data as an index indicating the magnitude of the fluctuation of the period of the peaks in the time series data. Then, if the magnitude of the fluctuation of the period of the peaks in the time series data is equal to or less than a threshold value, the determination unit 11 determines that the time series data has property a. On the other hand, if the magnitude of the fluctuation of the period of the peaks in the time series data exceeds the threshold value, the determination unit 11 determines that the time series data has property b (embodiment 1).

In the second example, the determination unit 11 performs a Fourier transform on the time series data to convert it into a spectrum. The determination unit 11 calculates the peak intensities of the spectrum. Then, if all peak intensities of the spectrum are equal to or greater than the threshold, the determination unit 11 determines that the time series data has property a. On the other hand, if one or more peak intensities of the time series data are below the threshold, the determination unit 11 determines that the time series data has property b (embodiment 2). Note that the method by which the determination unit 11 determines the property of the time series data is not limited to the first and second examples described here.

The determination unit 11 outputs the determination result of the properties of the time series data to the selection unit 12. The determination unit 11 also outputs the time series data to the identification unit 13.

The selection unit 12 selects a method for analyzing the time series data based on the characteristics of the time series data. The selection unit 12 is an example of a selection means.

In one example, the selection unit 12 receives a judgment result of the property of the time series data from the judgment unit 11. The selection unit 12 selects a method for analyzing the time series data based on the judgment result of the property of the time series data. For example, when the time series data has property a, the selection unit 12 selects a first method using nonnegative matrix factorization (NMF). In the method using nonnegative matrix factorization (hereinafter, NMF), a spectrogram in which the spectrum of the time series data is arranged in chronological order is decomposed into a basis matrix and an activation matrix. The activation matrix obtained in this way is the feature in the first method.

On the other hand, when the time series data has property b, the selection unit 12 selects the second method using Mel-Frequency Cepstrum Coefficients (MFCC). Mel-Frequency Cepstrum Coefficients (hereinafter, MFCC) are low-order components of a weighted cepstrum obtained by cepstrum analysis of the time series data. The MFCC obtained in this manner is a feature in the second method. The selection unit 12 notifies the identification unit 13 of the method (the first method or the second method) for analyzing the time series data.

The identification unit 13 uses the selected method to analyze the time series data and identify abnormal sounds contained in the time series data. The identification unit 13 is an example of an identification means.

In one example, the identification unit 13 receives time series data from the determination unit 11. The identification unit 13 is also notified of a method (either the first method or the second method) for analyzing the time series data from the selection unit 12. The identification unit 13 analyzes the time series data using the method selected by the selection unit 12. For example, when the first method is selected, the identification unit 13 first converts the time series data into a spectrogram. Then, the identification unit 13 obtains an activation matrix by decomposing the spectrogram using NMF. The identification unit 13 inputs the obtained activation matrix as a feature to a classifier (hereinafter referred to as classifier A) that has been machine-learned using the activation matrix as a feature. The classifier A classifies the time series data based on the feature of the input activation matrix, and outputs the classification result.

On the other hand, when the second method is selected, the identification unit 13 first obtains the MFCC by cepstral analysis of the time series data. The identification unit 13 inputs the MFCC obtained by the cepstral analysis as a feature to a classifier (hereinafter referred to as classifier B) that has been machine-learned using the MFCC as a feature. The classifier B classifies the time series data based on the input MFCC feature and outputs the classification result. In this way, the identification unit 13 uses classifier A or classifier B to classify the time series data according to the method selected by the selection unit 12. The identification unit 13 may output the classification result of the time series data to a subsequent processing unit (not shown), or may provide it to a recording medium or an external device.

(Operation of the abnormal sound identification device 10)
The operation of the abnormal sound identification device 10 according to the first embodiment will be described with reference to Fig. 2. Fig. 2 is a flowchart showing the flow of processes executed by each unit of the abnormal sound identification device 10.

First, time series data is input to the abnormal sound identification device 10. The time series data is an acoustic signal generated by collecting sounds emitted by equipment or parts, for example, in a moving train, in a factory, in the engine room of an automobile, etc., using a microphone. The abnormal sound identification device 10 receives time series data such as acoustic signals via any wireless or wired network. The abnormal sound identification device 10 then begins the following operations.

As shown in FIG. 2, the judgment unit 11 judges the property of the time series data (S1). In one example, the judgment unit 11 measures the time width (period) from one peak to the next peak of the time series data. Then, if the magnitude of the fluctuation in the period is equal to or less than the threshold, the judgment unit 11 judges that the time series data has property a. On the other hand, if the magnitude of the fluctuation in the period exceeds the threshold, the judgment unit 11 judges that the time series data has property b. The judgment unit 11 outputs the judgment result of the property of the time series data to the selection unit 12. The judgment unit 11 also outputs the time series data to the identification unit 13.

Next, the selection unit 12 selects a method for analyzing the time series data based on the properties of the time series data (S2). For example, if the time series data has property a, the selection unit 12 selects a first method using NMF. Thereafter, if the time series data has property b, the selection unit 12 selects a second method using MFCC. The selection unit 12 notifies the identification unit 13 of the method for analyzing the time series data.

The identification unit 13 identifies abnormal sounds contained in the time series data by analyzing the time series data using the method selected by the selection unit 12 (S3).

This completes the operation of the abnormal sound identification device 10 according to this embodiment 1.

(Effects of this embodiment)
According to the configuration of this embodiment, the determination unit 11 determines the properties of the time series data. The selection unit 12 selects a method for analyzing the time series data based on the properties of the time series data. The identification unit 13 analyzes the time series data using the selected method to identify abnormal sounds included in the time series data. The time series data includes various types of sounds (including abnormal sounds) and noises, and the properties of the time series data are also various. For example, the time series data may include abnormal sounds with large periodic fluctuations, or the noise may be loud (the target sound may be quiet).

The abnormal sound identification device 10 first determines the nature of the time series data, and then selects a method for analyzing the time series data based on the result of that determination. This allows the device to accurately identify abnormal sounds from time series data of various characteristics.

[Embodiment 2]
A second embodiment will be described with reference to Figures 3 to 5. In this second embodiment, an example of a method for determining the properties of time-series data will be described. In this second embodiment, the description of the first embodiment will be cited with respect to the configuration described in the first embodiment, and the description thereof will be omitted.

(Abnormal sound identification device 20)
Fig. 3 is a block diagram showing the configuration of an abnormal sound identification device 20 according to the second embodiment. As shown in Fig. 3, the abnormal sound identification device 20 includes a determination unit 21, a selection unit 12, and a recognition unit 13. The determination unit 21 of the abnormal sound identification device 20 includes a peak detection unit 24. The peak detection unit 24 detects peaks in the time series data.

An example of a method for determining the properties of time series data will be specifically described with reference to Figure 4. Figure 4 illustrates time series data having property a and time series data having property b. In Figure 4, the peaks detected by the peak detection unit 24 are indicated by dots (black circles).

In this embodiment 2, the judgment unit 21 judges the nature of the time series data based on the time width (called the period) from when a peak is detected in the time series data to when the next peak is detected.

In Figure 4, the period of peaks in the time series data is represented by the distance between the points indicating the peaks of the time series data (i.e., the length of the double-headed arrows). In the time series data on the top, the period of peaks in the time series data is almost constant. In other words, the time series data on the top has a small difference in period (periodic fluctuation). On the other hand, in the time series data on the bottom, there is variation in the period of peaks in the time series data. In other words, the time series data on the bottom has a large difference in period (periodic fluctuation).

The determination unit 21 compares the magnitude of the difference in the period of the peaks in the time series data (periodic fluctuation) with a predetermined threshold. For example, when the magnitude of the fluctuation in the period of the peaks in the time series data is expressed as the deviation of the difference, the threshold X is 0.5. In this example, when the fluctuation in the period of the peaks in the time series data is X = 0.5 or less, the determination unit 21 determines that the time series data has property a. On the other hand, when the fluctuation in the period of the peaks in the time series data exceeds X = 0.5, the determination unit 21 determines that the time series data has property b.

(Operation of the abnormal sound identification device 20: S1)
The operation of the abnormal sound identification device 20 according to the second embodiment will be described with reference to Fig. 5. Here, only the details of the flow of the process executed by the determination unit 21, that is, the content of step S1 shown in Fig. 2 will be described.

As in the first embodiment, the abnormal sound identification device 20 receives the time series data. The determination unit 21 of the abnormal sound identification device 20 then determines the nature of the time series data as described below.

As shown in FIG. 5, the peak detection unit 24 of the judgment unit 21 detects a peak in the time series data (S21).

The judgment unit 21 calculates the fluctuation in the period of the peaks in the time series data based on the time width between the peaks in the time series data (S22).

The judgment unit 21 judges whether the magnitude of fluctuation in the period of the peaks in the time series data is below a threshold value (S23).

If the magnitude of the fluctuation in the period of the peaks in the time series data is equal to or less than the threshold (Yes in S23), the determination unit 21 determines that the time series data has property a (S24A). On the other hand, if the magnitude of the fluctuation in the period of the peaks in the time series data exceeds the threshold (No in S23), the determination unit 21 determines that the time series data has property b (S24B).

This ends the processing of the determination unit 21. After that, the processing proceeds to the processing (step S2) of the selection unit 12 described in the first embodiment. In the second embodiment, the description of the processing after the processing (step S2) of the selection unit 12 is omitted.

(Effects of this embodiment)
According to the configuration of this embodiment, the determination unit 21 determines the properties of the time series data. The selection unit 12 selects a method for analyzing the time series data based on the properties of the time series data. The identification unit 13 analyzes the time series data using the selected method to identify abnormal sounds included in the time series data. The time series data includes various types of sounds (including abnormal sounds) and noises, and the properties of the time series data are also various. For example, the time series data may include abnormal sounds with large periodic fluctuations, or the noise may be loud (the target sound may be quiet).

The abnormal sound identification device 10 first determines the nature of the time series data, and then selects a method for analyzing the time series data based on the result of that determination. This allows the device to accurately identify abnormal sounds from time series data of various characteristics.

Furthermore, according to the configuration of the second embodiment, the determination unit 21 includes a peak detection unit 24 that detects peaks in the time series data. The determination unit 21 determines the nature of the time series data based on the time width from when a peak in the time series data is detected until the next peak is detected. The magnitude of fluctuation in the period of the peaks in the time series data can be calculated from the time width from when a peak in the time series data is detected until the next peak is detected. Then, by comparing the magnitude of fluctuation in the period of the peaks in the time series data with a threshold value, it is possible to determine whether the nature of the time series data has a relatively small fluctuation in the period of the peaks or a relatively large fluctuation in the period of the peaks in the time series data.

[Embodiment 3]
A third embodiment will be described with reference to Figures 6 to 9. In this third embodiment, another example of a method for determining the properties of time-series data will be described. In this third embodiment, the description of the first embodiment will be cited with respect to the configuration described in the first embodiment, and the description thereof will be omitted.

(Abnormal sound identification device 30)
Fig. 6 is a block diagram showing the configuration of an abnormal sound identification device 30 according to the third embodiment. As shown in Fig. 6, the abnormal sound identification device 30 includes a determination unit 31, a selection unit 12, and a recognition unit 13. The determination unit 31 of the abnormal sound identification device 30 includes a data conversion unit 34. The data conversion unit 34 converts a time domain signal such as time series data or a waveform into a frequency domain signal such as a spectrum or a spectrogram. An example of converting time series data into a spectrogram will be described below.

Figure 7 shows an example of a spectrogram converted from time series data. In the spectrogram shown in Figure 7, the frequency spectrum intensity is represented by shading. Furthermore, the peaks of the frequency spectrum are shown by thick lines (bars). In the example shown in Figure 7, some of the peaks of the frequency spectrum are inclined with respect to the vertical and horizontal axes. This indicates that the peak frequency is shifting over time. In other words, the period of the peaks (= 1/peak frequency) in the original time series data is fluctuating.

Figure 8 is a graph showing an example of a frequency spectrum converted from time series data. The frequency spectrum corresponds to a given time width in the spectrogram. In Figure 8, the peaks of the frequency spectrum are shown as dots (black circles) on the graph.

A sharp, high peak in the frequency spectrum corresponds to a nearly constant period of peaks in the original time series data over a given time span (i.e., small periodic fluctuations). On the other hand, a dull, low peak in the frequency spectrum corresponds to a variance in the period of peaks in the original time series data over a given time span (i.e., large periodic fluctuations).

The determination unit 31 determines the nature of the time series data based on the peak intensity of the frequency spectrum extracted from the spectrogram for each predetermined time width. For example, the determination unit 31 calculates the difference between the peak intensity in the frequency spectrum and the average intensity in a predetermined band centered on the peak frequency. The determination unit 31 compares the obtained difference with a threshold Y. In this example, if the difference between the peak intensity in the frequency spectrum and the average intensity in a predetermined band centered on the peak frequency is equal to or greater than the threshold Y, the determination unit 31 determines that the time series data has the property a. On the other hand, if the difference between the peak intensity in the frequency spectrum and the average intensity in a predetermined time width is less than the threshold Y, the determination unit 31 determines that the time series data has the property b.

In addition, by feeding back information on the reliability of the abnormal sound identification result by the downstream identification unit 13 to the judgment unit 31, the judgment unit 31 may update the threshold value Y so as to increase the reliability of the abnormal sound identification result by the identification unit 13.

(Operation of the abnormal sound identification device 30: S1)
The operation of the abnormal sound identification device 30 according to the third embodiment will be described with reference to Fig. 9. Here, only the details of the flow of the process executed by the determination unit 31, that is, the content of step S1 shown in Fig. 2 will be described.

As in the first embodiment, the abnormal sound identification device 30 receives the time series data. The determination unit 31 of the abnormal sound identification device 30 then determines the nature of the time series data as described below.

As shown in Figure 9, the data conversion unit 34 of the judgment unit 31 converts the time series data (time domain signal) into a spectrogram (Figure 7) (frequency domain signal) (S31).

The determination unit 31 generates a frequency spectrum for each segment by extracting a segment of a predetermined time width from the spectrogram. The determination unit 31 calculates the peak intensity in the frequency spectrum (S32).

The determination unit 31 determines whether the peak intensity in the frequency spectrum is equal to or greater than a threshold (S33). For example, the threshold is the average intensity in a predetermined band centered on the peak frequency.

If the peak intensity in the frequency spectrum is equal to or greater than the threshold (Yes in S33), the determination unit 31 determines that the time series data has property a (S34A). On the other hand, if the peak intensity in the frequency spectrum is less than the threshold (No in S33), the determination unit 31 determines that the time series data has property b (S34B).

This completes the processing of the determination unit 31. After that, the processing proceeds to step S2 of the selection unit 12 described in the first embodiment. In the third embodiment, the processing subsequent to step S2 of the selection unit 12 will not be described.

(Method of determining threshold value)
Here, a configuration has been described in which the determination unit 31 determines the property of the time series data by comparing the peak intensity of the frequency spectrum with a threshold value. Here, an example of a method for determining the threshold value of the peak intensity of the frequency spectrum will be described.

Figure 10 is an example graph showing the distribution of scores used to determine the peak intensity threshold of a frequency spectrum.

The determination unit 31 calculates scores for a large number of learning data that contain the same or almost the same number of both time series data that have been determined to have property a and time series data that have been determined to have property b. The score here is the difference between the average intensity in a predetermined band centered on the peak frequency and the peak intensity. From the score calculation result, a distribution of scores as shown in FIG. 10 is obtained. Then, the determination unit 31 determines a threshold value so that time series data having property a and time series data having property b can be distinguished based on the distribution of scores. For example, the determination unit 31 determines a threshold value that is twice the maximum score of time series data having property b. If the score of a certain time series data is equal to or greater than the threshold value, the time series data is highly likely to have property a, while if the score of a certain time series data is below the threshold value, the time series data is highly likely to have property b. The determination unit 31 can determine the property of the time series data as described above by using the threshold value determined in this manner.

(Effects of this embodiment)
According to the configuration of this embodiment, the determination unit 31 determines the properties of the time series data. The selection unit 12 selects a method for analyzing the time series data based on the properties of the time series data. The identification unit 13 analyzes the time series data using the selected method to identify abnormal sounds contained in the time series data. The time series data includes various types of sounds (including abnormal sounds) and noises, and the properties of the time series data are also various. For example, the time series data may include abnormal sounds with large periodic fluctuations, or the noise may be loud (the target sound may be quiet).

The abnormal sound identification device 10 first determines the nature of the time series data, and then selects a method for analyzing the time series data based on the result of that determination. This allows the device to accurately identify abnormal sounds from time series data of various characteristics.

Furthermore, according to the configuration of the third embodiment, the determination unit 31 includes a data conversion unit 34 that converts the time series data into a spectrogram. The determination unit 31 determines the nature of the time series data based on the peak intensity of the frequency spectrum cut out from the spectrogram for each predetermined time width. A sharp and strong peak in the frequency spectrum corresponds to a small fluctuation in the period of the peak in the original time series data. Then, by comparing the peak intensity of the frequency spectrum with a threshold value, it is possible to determine whether the nature of the time series data has a relatively small fluctuation in the period of the peak or a relatively large fluctuation in the period of the peak in the time series data.

(Modification)
In a modification of any of the first to third embodiments, the identification unit 13 uses three or more classifiers to identify abnormal sounds contained in the time-series data.

For example, the identification unit 13 in this modified example uses a classifier B that has been machine-learned using MFCC as a feature, as well as a classifier (hereinafter referred to as classifier C) that uses DCTC (Discrete Cosine Transform Coefficients) as a feature. The identification unit 13 identifies abnormal sounds contained in the time-series data using each of the two classifiers and compares the reliability of the identification results. If the reliability of the identification result by classifier B is higher, the identification unit 13 outputs the identification result by classifier B. On the other hand, if the reliability of the identification result by classifier C is higher, the identification unit 13 outputs the identification result by classifier C.

In another variation, the identification unit 13 may use multiple classifiers depending on the location where the acoustic signal that is the source of the time series data was acquired. In this variation, each classifier is previously associated with information indicating a different location. The identification unit 13 in this variation receives information indicating a location linked to the time series data from the determination unit 11 along with the time series data. The identification unit 13 selects a corresponding classifier from the multiple classifiers based on the information indicating the location. Then, the identification unit 13 uses the selected classifier to identify abnormal sounds included in the time series data.

According to the configuration of this modified example, one of the two classifiers is selected, which has a higher reliability of the classification result, thereby improving the reliability of the classification result output by the classification unit 13.

[Hardware configuration]
Each of the components of the abnormal sound identification devices 10, 20, and 30 described in the first to third embodiments is represented by a functional block. Some or all of these components are realized by an information processing device 900 as shown in Fig. 11. Fig. 11 is a block diagram showing an example of the hardware configuration of the information processing device 900.

As shown in FIG. 11, the information processing device 900 includes, as an example, the following configuration.

・CPU (Central Processing Unit) 901
・ROM (Read Only Memory) 902
・RAM (Random Access Memory) 903
Program 904 loaded into RAM 903
A storage device 905 for storing the program 904
A drive device 907 for reading and writing data from and to the recording medium 906
A communication interface 908 for connecting to a communication network 909
An input/output interface 910 for inputting and outputting data
A bus 911 connecting each component
Each of the components of the abnormal sound identification devices 10, 20, and 30 described in the first to third embodiments is realized by the CPU 901 reading and executing a program 904 that realizes these functions. The program 904 that realizes the functions of each component is, for example, stored in advance in the storage device 905 or the ROM 902, and is loaded into the RAM 903 and executed by the CPU 901 as necessary. The program 904 may be supplied to the CPU 901 via the communication network 909, or may be stored in advance in the recording medium 906, and the drive device 907 may read out the program and supply it to the CPU 901.

According to the above configuration, the abnormal sound identification devices 10, 20, and 30 described in the first to third embodiments are realized as hardware. Therefore, the same effects as those described in the first to third embodiments can be achieved.

[Additional Notes]
One embodiment of the present invention is described as follows, but is not limited to the following.

(Appendix 1)
A determination means for determining the nature of the time series data;
A selection means for selecting a technique for analyzing the time series data based on the characteristics of the time series data;
and an identification means for identifying an abnormal sound contained in the time-series data by analyzing the time-series data using the selected method.

(Appendix 2)
The determination means includes a peak detection means for detecting a peak in the time series data,
The data analysis device according to claim 1, wherein the determination means determines the property of the time series data based on a time width from when the peak of the time series data is detected to when the next peak is detected.

(Appendix 3)
The determination means includes a data conversion means for converting the time series data into a spectrogram,
The data analysis device according to claim 1, wherein the determining means determines the property of the time series data based on peak intensities of frequency spectra extracted from the spectrogram for each predetermined time width.

(Appendix 4)
The determination means determines a magnitude of variation in a period of a periodic component contained in the time series data,
The data analysis apparatus according to any one of claims 1 to 3, wherein the selection means selects a method from a plurality of methods for analyzing the time series data according to a magnitude of variation in the period.

(Appendix 5)
The data analysis apparatus according to claim 4, wherein the selection means selects Nonnegative Matrix Factorization (NMF) when the magnitude of the variation in the period is below a threshold value.

(Appendix 6)
Determine the nature of the time series data,
selecting a technique for analyzing the time series data based on the characteristics of the time series data;
A data analysis method for identifying abnormal sounds contained in time-series data by analyzing the time-series data using the selected technique.

(Appendix 7)
Determining a nature of the time series data;
selecting a technique for analyzing the time series data based on the characteristics of the time series data;
and identifying an abnormal sound contained in the time-series data by analyzing the time-series data using the selected technique.

The present invention can be used, for example, in an abnormal sound identification device that identifies abnormal sounds emitted by equipment or parts in railways, automobile engine compartments, factories, and other places.

REFERENCE SIGNS LIST 10 Abnormal sound identification device 11 Determination unit 12 Selection unit 13 Identification unit 20 Abnormal sound identification device 24 Peak detection unit 30 Abnormal sound identification device 34 Data conversion unit

Claims (6)

A determination means for determining the nature of the time series data;
A selection means for selecting a technique for analyzing the time series data based on the characteristics of the time series data;
and an identification means for identifying an abnormal sound included in the time-series data by analyzing the time-series data using the selected technique ,
The determination means determines a magnitude of variation in a period of a periodic component contained in the time series data,
The selection means selects a method according to the magnitude of the variation in the period from among a plurality of methods for analyzing the time series data.
Data analysis equipment. The determination means includes a peak detection means for detecting a peak in the time series data,
The data analysis apparatus according to claim 1 , wherein the determining means determines the property of the time series data based on a time width from when the peak of the time series data is detected to when the next peak is detected. The determination means includes a data conversion means for converting the time series data into a spectrogram,
The data analysis apparatus according to claim 1 , wherein the determining means determines the property of the time series data based on peak intensities of frequency spectra extracted from the spectrogram for each predetermined time width. 4. The data analysis apparatus according to claim 1 , wherein the selection means selects NMF (Nonnegative Matrix Factorization) when the magnitude of the variation in the period is below a threshold value. The computer
Determine the nature of the time series data,
selecting a technique for analyzing the time series data based on the characteristics of the time series data;
A data analysis method for identifying an abnormal sound included in time-series data by analyzing the time-series data using the selected technique,
The computer,
determining a magnitude of variation in a period of a periodic component contained in the time series data;
Selecting a method according to the magnitude of the variation in the period from among a plurality of methods for analyzing the time series data.
Data analysis methods. Determining a nature of the time series data;
selecting a technique for analyzing the time series data based on the characteristics of the time series data;
Analyzing the time-series data using the selected technique to identify abnormal sounds included in the time-series data ;
A program for causing a computer to execute the above,
determining a magnitude of variation in a period of a periodic component contained in the time series data;
selecting a method according to the magnitude of the variation in the period from among a plurality of methods for analyzing the time series data;
A program for causing the computer to execute the above .

Images