How should different numbers of explanatory variables be handled by row in machine learning? I have a problem while making a predictive model, so I’m leaving a question.