Question

Write out the steps for one model selection technique (backward/forward-elimination with the R-squared, backward/forward-elimination with the p-value approach, etc.).

Answer 1

Backward/Forward- elimination with the p-value approach:

The p-value is the probability that is calculated by assuming that the null hypothesis is true.
1. Under right trial test, p-value is the area to the right of test statistic value.

2. Under left trail test, p-value is the area to the left of tge test statistic value.

3. Under two tailed test;

• P- value = 2 (area to the right of the Test statistic value if and only iff T.S. is positive.)
• p-value = 2 ( area to the left of the test statistic value if and only iff T.S. is negative.)

Decision based on p-value-

The smaller the p-value, the more evident it would be that the sample is against the null hypothesis and for the alternative hypothesis.

When we select a significance level, 'alpha' (a) , then reject hypothesis if p-value is less than the value of alpha (a) or do not reject the hypothesis if the p-value is more than the value of alpha (a).

For a backward elimination, we will continue the steps until the condition holds false.