PCA
From Ufldl
(→Recovering an Approximation of the Data) |
(→Number of components to retain) |
||
| Line 239: | Line 239: | ||
the variance will also be a much more easily interpretable description than saying | the variance will also be a much more easily interpretable description than saying | ||
that you retained 120 (or whatever other number of) components. | that you retained 120 (or whatever other number of) components. | ||
| + | |||
| + | == What works well == | ||
For PCA to work, usually we want each of the features <math>\textstyle x_1, x_2, \ldots, x_n</math> | For PCA to work, usually we want each of the features <math>\textstyle x_1, x_2, \ldots, x_n</math> | ||
| Line 292: | Line 294: | ||
and that <math>\textstyle \mu^{(i)}</math> here is the mean intensity of the image <math>\textstyle x^{(i)}</math>. In particular, | and that <math>\textstyle \mu^{(i)}</math> here is the mean intensity of the image <math>\textstyle x^{(i)}</math>. In particular, | ||
this is not the same thing as estimating a mean value separately for each pixel <math>\textstyle x_j</math>. | this is not the same thing as estimating a mean value separately for each pixel <math>\textstyle x_j</math>. | ||
| + | |||
| + | == Non-natural images == | ||
If you are training your algorithm on images other than natural images (for | If you are training your algorithm on images other than natural images (for | ||
| Line 360: | Line 364: | ||
when training on natural images, using the per-image mean normalization | when training on natural images, using the per-image mean normalization | ||
as in Equations~(\ref{eqn-normalize1}-\ref{eqn-normalize2}) | as in Equations~(\ref{eqn-normalize1}-\ref{eqn-normalize2}) | ||
| - | would be a reasonable default. | + | would be a reasonable default. |
== PCA on Images == | == PCA on Images == | ||
