R-Squared (R²)

INFO

Represents the proportion of variance explained by a regression model, assessing its fit.

How It Works

$$R^2=1-\frac{\sum (y_i-{\hat{y}}_i{)}^2}{\sum (y_i-\bar{y}{)}^2}$$

• $\bar{y}$: Mean of actual values
• $R^2$ ranges from 0 to 1 (or negative if model is worse than baseline)

What to Look For

• Higher $R^2$ = better fit
• Can be misleading with non-linear models or overfitting
• Use alongside residual plots for deeper insight

Application Models

• Deep Neural Network (DNN)
• Recurrent Neural Network (RNN)