Model Evaluation Dashboard

Among all 6,000 data samples from the generated dataset, 2,000 samples were split as the test data. These samples had not been seen by the trained models, and this approach was used for evaluating and validating the models.

Multi-Output Regression Plots

heating

R^2: 0.3922

cooling

R^2: 0.5790

lighting

R^2: 0.9993

equipments

R^2: 0.9371

Neural Network Plots

accuracy of energy grade prediction

accuracy: 0.89

Decision Tree Plots

confusion matrix

scores for class energy grade

class 0, 1, and 2 represents the EC, EC+, and EC++ respectively.

learnin curve of model

how much the importance of inputs