Where, a is the Y intercept of the regression line and b is the slope of the regression line.
E(Y) is expected value for given X.
Now, for linear regression SSR, SSE and SST are defined as,
SST = total sum of squares = ∑(yi – ӯ)2
SSR = sum of squares due to regression= ∑ (ŷi – ӯ)2
SSE = sum of squares due to error = ∑ (yi – ŷi)2
SST = SSR +SSE
The coefficient of determination is r2 = SSR/SST
SSR = sum of squares due to regression
SST = total sum of squares
The value of co-efficient of determination varies between 0 and 1. The correlation is very strong the value of co-efficient will be near to one. If the value is near to zero, the regression model isn’t good enough to describe the data set.