Skip to content

Latest commit

 

History

History
30 lines (14 loc) · 659 Bytes

README.md

File metadata and controls

30 lines (14 loc) · 659 Bytes

SVM

SVM classification work on dataset

kernel = linear:

accuracy = 0.91 - 0.97

kernel = rbf:

accuracy = 1.0

Above performance is without PCA, means all 13 features is used.

Implmentation of result:

applying PCA on dataset, using SVM to get optimization

SVM with linear kernel:

img

SVM with rbf kernel:

Get performance that is not that good. We can see that PCA influences the performance very much. Even the best performance with rbf kernel couldn't reach good performance by the principal component.

img