Skip to content

Latest commit

 

History

History
10 lines (9 loc) · 673 Bytes

LDA_Dimension_Reduction.md

File metadata and controls

10 lines (9 loc) · 673 Bytes
Raw

Linear Discriminant Analysis (LDA)

  1. LDA helps in identifying and finding patterns to reduce the dimensions of the dataset such that the separation between the various classes in maximum.
  2. More theory can be studied from LDA
  3. Ways to compute eigen values, eigen vectors - Eigen Value Decomposition

Key Points

  1. LDA is very similar to PCA, but if you need to reduce the dimensions and also keep good separation between the classes then LDA would be good.
  2. Draw back of LDA is that it works only for supervised classification tasks only.