The analysis explores the model’s behavior by varying income, job type, and city. Notably, the model’s predictions are predominantly influenced by income, revealing a clear positive correlation. As income increases, energy consumption tends to rise. On the other hand, simulating job types and cities yields relatively consistent results, suggesting that the model’s predictions are less sensitive to these categorical features. This indicates that income is the most significant factor in predicting energy consumption, while job type and city have a relatively minor impact.
In conclusion, the Random Forest Regressor model demonstrates robust predictive capabilities for estimating energy consumption. It achieves good accuracy in its predictions, as evidenced by low MAE, MSE, and a high R2 score. The model’s reliance on income as the dominant predictor implies that changes in income strongly influence energy consumption. However, the model appears less sensitive to variations in job type and city. These insights can help stakeholders better understand the dynamics of energy consumption and make informed decisions for resource allocation and policy planning.