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

Table 5

Feature importance based on the RF model.

Variable code	Variable importance (% IncMSE)
Usage_Type	0.4000
Area	0.1600
Mean_Station_Pressure	0.0800
Industrial_Labor	0.0800
Household	0.0400
Mean_Relative_Humidity	0.0300
Mean_Maximum_Temperature	0.0200
Mean_Minimum_Temperature	0.0200
Mean_Drybulb_Temperature	0.0200
Expenditure	0.0200
Income	0.0200
Season	0.0100
Mean_MSL_Pressure	0.0100
Total_Rainfall	0.0100
Agriculturists	0.0100
Industrial_Plant	0.0100
Liabilities	0.0100
Population_Ratio	0.0100
TOU	0.0000
Population_N	0.0000
Electrical_Substation	0.0000