Data generation apparatus, data generation method, and non-transitory computer-readable storage medium for storing program

1. A data generation method comprising:

executing a first generation process that includes generating a Betti number series corresponding to a contribution rate by performing persistent homology processing on a first point set, the first point set being generated by using a plurality of pieces of time series data and the contribution rate of each of the plurality of pieces of time series data, each of points included in the first point set being represented by coordinates;

executing a second generation process that includes generating a characteristic image from a plurality of the Betti number series, the plurality of Betti number series being generated by performing the first generation process on each of the plurality of contribution rates; and

executing a third generation process that includes generating machine learning data in which the characteristic image and a classification corresponding to the plurality of pieces of time series data are associated with each other.

2. The data generation method according to claim 1 , further comprising:

executing a fourth generation process that includes generating first data based on the generated machine learning data, the first data being data on a relationship between relevance and the contribution rate, the relevance indicating a degree of influence on a specific classification among a plurality of the classifications; and

executing a fifth generation process that includes generating second data on the specific classification based on the first data, the second data being data on a correlation between the plurality of pieces of time series data.

3. The data generation method according to claim 2 ,

wherein the machine learning data is machine learning data based on a convolutional neural network, and

wherein the fourth generation process is configured to

calculate relevance on the specific classification, for each of values of the plurality of contribution rates, based on absolute values of elements of a weighting matrix between a last characteristic layer and an identification layer contained in the machine learning data.

4. The data generation method according to claim 1 ,

wherein the machine learning data is machine learning data based on a single-layer learning model.

5. The data generation method according to claim 2 ,

wherein the fifth generation process is configured to

generate a plurality of pieces of relationship data based on a maximum value and a minimum value of relevance contained in the first data as well as positions of the maximum value and the minimum value, the plurality of pieces of relationship data including any of first relationship data, second relationship data, third relationship data, and fourth relationship data, the first relationship data being data indicating that each of the plurality of pieces of time series data independently influences the specific classification, the second relationship data being data indicating that any of the plurality of pieces of time series data influences the specific classification, the third relationship data being data indicating that each of the plurality of pieces of time series data influences the specific classification and that the plurality of pieces of time series data have time correlation, the fourth relationship data being data indicating that each of the plurality of pieces of time series data influences the specific classification and that the plurality of pieces of time series data have no time correlation.

6. The data generation method according to claim 5 ,

wherein the fifth generation process is configured to

generate the first relationship data when a ratio of the maximum value to the minimum value is lower than or equal to a predetermined value,

generate the second relationship data when the maximum value and the minimum value are at ends of a column of relevance contained in the first data,

generate the third relationship data when the maximum value is at a middle of the column of relevance contained in the first data, and

generate the fourth relationship data when the minimum value is at a middle of the column of relevance contained in the first data.

7. The data generation method according to claim 1 ,

wherein the plurality of pieces of time series data are a plurality of pieces of normalized time series data.

8. An apparatus of data generation, the apparatus comprising:

a memory; and

a processor coupled to the memory, the processor being configured to

execute a first generation process that includes generating a Betti number series corresponding to a contribution rate by performing persistent homology processing on a first point set, the first point set being generated by using a plurality of pieces of time series data and the contribution rate of each of the plurality of pieces of time series data, each of points included in the first point set being represented by coordinates;

execute a second generation process that includes generating a characteristic image from a plurality of the Betti number series, the plurality of Betti number series being generated by performing the first generation process on each of the plurality of contribution rates; and

execute a third generation process that includes generating machine learning data in which the characteristic image and a classification corresponding to the plurality of pieces of time series data are associated with each other.

9. A non-transitory computer-readable storage medium storing a program which causes a processor to perform processing of data generation, the processing comprising:

executing a first generation process that includes generating a Betti number series corresponding to a contribution rate by performing persistent homology processing on a first point set, the first point set being generated by using a plurality of pieces of time series data and the contribution rate of each of the plurality of pieces of time series data, each of points included in the first point set being represented by coordinates;

executing a second generation process that includes generating a characteristic image from a plurality of the Betti number series, the plurality of Betti number series being generated by performing the first generation process on each of the plurality of contribution rates; and

executing a third generation process that includes generating machine learning data in which the characteristic image and a classification corresponding to the plurality of pieces of time series data are associated with each other.