0

I have an data sheet with almost 2000 input parameters and 4 output parameters. I am to optimize the input parameters to define the output

I am not sure whether the input parameters are linearly related so I am looking for a non linear dimensionality reduction algorithm. Will Self organizing maps be a good option for reducing the 2000 input parameters to few components.

I did try PCA but it is not quite helpful. Please if anyone could suggest any standard algorithms and codes that I could try?

shr m
  • 27
  • 1
  • What is your task? classification? – Saeed Oct 11 '15 at 17:57
  • yes! i am trying to reduce the number of input parameters so as to best describe the output..type of classification with optimzing the input parameters – shr m Oct 11 '15 at 21:02
  • 1
    See LDA dimenlionality reduction. Google it and you'll find useful info. – Saeed Oct 11 '15 at 22:18
  • is LDA a non linear dimensionality reduction algorithm? my input parameters are not linearly related... – shr m Oct 11 '15 at 23:04
  • 1
    See this link: http://tcts.fpms.ac.be/asr/project/sprach/report97/node164.html – Saeed Oct 12 '15 at 12:25

0 Answers0