How to classify using PCA ("pcares" function)

2 次查看（过去 30 天）

Kong 2020-3-28

Hello.

I want to do a video classification.

I found some function about PCA.

[residuals,reconstructed] = pcares(X,5)

After using pcares(X,5), Can I classify data using "reconstructed"?

I mean, "reconstructed" is 3600 x 25.

1. After flattening each data, I got 90000 X 1.

2. I have 90 data, so I got 90 X 90001 (the last column is class).

3. Using classification methods like SVM, KNN, etc.

4. calculate the accuracy of the classification.

Is it a reasonable approach? or do I need to use other methods as PCA not "pcares".

When I use "reconstructed" data after "pcares", the accuracy of classification is so low.

Please let me know how to deal with this issue.

在 Help Center 和 File Exchange 中查找有关 Dimensionality Reduction and Feature Extraction 的更多信息

Find the treasures in MATLAB Central and discover how the community can help you!

Translated by