Measure of fit
Summary
Setup
- Given an OLS trendline with intercept
\[y=b_1+b_2x\]
- where \(b_1,b_2\) are determined by the OLS formula and \(e_i, {\hat{y}}_i\) are the OLS residuals and the OLS fitted values
TSS, ESS, RSS
- Total sum of squares (TSS)
\[TSS=\sum_{i=1}^{n}{ {\left( y_i-\bar{y} \right)}^2 }\]
- Explained sum of squares (ESS)
\[ESS=\sum_{i=1}^{n}{ {\left( {\hat{y}}_i-\bar{y} \right)}^2 }\]
- Residual sum of squares ( RSS ) (as before)
\[RSS=\sum_{i=1}^{n}{ e_i^2 }\]
- Note:
- Sample variance in \(y\) -data is equal to \(TSS/(n-1)\)
- Sample variance in \(\hat{y}\) -data is equal to \(ESS/(n-1)\)
- Sample variance in \(e\) -data is equal to \(RSS/(n-1)\)
Coefficient of determination
- Result: \(TSS=ESS+RSS\)
- Coefficient of determination
\[R^2= \frac{ESS}{TSS}=1- \frac{RSS}{TSS}\]
- Result: \(0≤R^2≤1\)