Knowledge aided feature engineering

A computer program product to employ feature engineering training to improve classifier performance, the computer program product comprising:

• a computer readable storage medium having program code embodied therewith, the program code executable by a processor to:
• select a machine learning (ML) model with a learning program, the selected ML model operatively associated with a first structured dataset;
• identify a plurality of first features in the first structured data set directed at performance of the selected ML model;
• assess a second structured dataset with respect to the identified features in the first dataset, identify new features in the second structured dataset semantically related to the identified plurality of first features, and dynamically augment the first structured dataset with the identified new features in the second structured dataset;
• and
• apply the dynamically augmented first structured dataset to the selected ML model, including train an embedded learning algorithm of the selected ML model using the dynamically augmented first structured dataset.

A feature engineering training method for improving classifier performance, comprising:

• selecting a machine learning (ML) model with a learning program, the ML model operatively associated with a first structured dataset;
• identifying a plurality of first features in the first structured data set directed at performance of the selected ML model;
• assessing a second structured dataset with respect to the identified features in the first structured dataset, identifying new features in the second structured dataset semantically related to the identified plurality of first features, and dynamically augmenting the first structured dataset with the identified new features in the second structured dataset;
• and
• applying the dynamically augmented first structured dataset to the selected ML model, including training an embedded learning algorithm of the selected ML model using the dynamically augmented first structured dataset.