Data model configuration method for learning data, learning data generation apparatus, and machine learning method

A learning data generation apparatus that generates learning data for machine learning, the learning data generation apparatus comprising:

• a processor configured to execute a learning data generator, and a machine learning unit, wherein the learning data generator is configured to:
• prepare filter data, the filter data specifying a degree of abstraction or a degree of detail, the degree of abstraction or the degree of detail of the data items to be designated for each of the data items and sorting the data items into an objective variable and an explanatory variable, and
• configure, by an information processing device, a data model in which the data item to be the objective variable and the data item to be the explanatory variable are defined to extract a data item to be used for training data from a database using at least one filter that acts on the data items based on the filter data, enables specification of the degree of abstraction or the degree of detail of the data items for each data item, and allocates data items to the objective variable and the explanatory variable,
• wherein the machine learning unit is configured to train a machine learning model using the objective variable and the explanatory variable;
• wherein a plurality of databases are used as a basis of the learning data and the data item of each of the plurality of databases have a hierarchical structure of the degree of abstraction or the degree of detail,
• wherein the at least one filter is applied to each of the plurality of databases and functions as an integrated filter that extracts data to be used for the learning data from the plurality of databases and integrates the data,
• wherein the at least one filter applied to each of the plurality of databases has a different characteristic;
• and
• wherein the filter further determines whether the data item of each of the plurality of databases is not to be used.