Fig. 2

The process of supervised machine learning modelling. Following the feature engineering stage (a), once an appropriate feature set is determined, the next phase involves dividing the data into training and testing datasets (b). The training dataset is employed to construct the predictive model, while the testing dataset, which is not involved in the model building process, is reserved for assessing the expected predictive performance. In statistical terms, this process is analogous to drawing inferences about a population based on a finite and random sample. Given that the correct values for the target variables are known, it is possible to objectively verify the model’s performance (c)