12.4 Decomposition of the Sample Variation in Y
Let us express the variance of the random variable Y as
We noted above that there were two components to Yi—a systematic component reflecting the linear influence of Xi and a random component εi due to chance factors. A moment's reflection reveals that if most of the variation in Y is due to random factors, then the estimated least squares line is probably worthless as far as predictions are concerned. However, if most of the variation in Y is due to the linear influence of X, then we can obtain a meaningful regression equation. So the question arises—“How much of the variation in Y is due to the linear influence of X and how much can be attributed to random factors? We can answer this question with the aid of Fig. 12.7.
Figure 12.7 Decomposition of
into
and ei.
Since our goal is to explain the variation in Y, let's start with the numerator of Equation (12.10). That is, from Fig. 12.7,
where is attributed to the linear ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access