Hybrid Machine Learning Model for Electricity Consumption Prediction Using Random Forest and Artificial Neural Networks

Table 2

Sampling procedures used for modeling.

Sample	No. of cases and % in the subsample	10-fold cross-validation
Training	619,154 cases (70%)	Train: 557,238 cases (90% of learning subsample); cross-validation: 61,915 cases (10% of learning subsample)
Test	265,352 cases (30%)
Total	884,505 cases (100%)