Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Consider Fisher's Linear Discriminant Analysis (LDA).

Let $\mu_0 \in \mathcal{R}^D$ and $\mu_1 \in \mathcal{R}^D$ be means of the two classes. Similarly, let $\Sigma_0 \in \mathcal{R}^{D\times D}$ and $\Sigma_1 \in \mathcal{R}^{D\times D}$ be the covariance matrices.

LDA considers the problem of finding a vector $w \in \mathcal{R}^D$ such that the following statistic is maximized:

$$S = \frac{(w^T\mu_0 - w^t\mu_1)^2}{w^T(\Sigma_0+\Sigma_1)w}$$

Does the dimension $D$ affect the above statistic? In practical terms, should I be comfortable comparing two values, $S_a$ and $S_b$, obtained from data in dimensions $D_a$ and $D_b \neq D_a$?

share|cite|improve this question

According to wikipedia the score $S$ is a ratio of between-class variance to within-class variance. This doesn't seem unreasonable to compare between different dimension sizes; regardless of dimension size $S \gt 1$ will mean more between-class than within-class variance.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.