Interpretation of coefficient of determination
If computing a causal linear regression model of y = a + bX and the resultant r2 is very near zero, then one would be able to conclude that:
1. y = a + bX is a good forecasting method.
2. Y = a + bX is not a good forecasting method.
3. a multiple linear regression model is a good forecasting method for the data.
4. a multiple linear regression model is not a good forecasting method for the data.