I have binary data (0,1) with set dimensions (for example 400). My data matrix is 10000*400. I need to reduce dimensions for the binary data (100 dimension). Can I use PCA for the data?
When I use PCA, my data in new space will not be binary!
You can use rough set theory coupled with any search method (e.g. evolutionary algorithms) to calculate the reducts of the system, i.e. the set of indispensable attributes, as shown in the paper below.
Chapter Rough Sets and Evolutionary Computation to Solve the Feature...
Define a similarity measure (like Tanimoto), calculate the distance matrix, and use distance preserving mapping, like Sammon projection or Multi Dimensional Scaling. (for implementation you can use my toolbox: http://www.abonyilab.com/software-and-data/fclusttoolbox)
Thank you Rafael .
My problem is not feature selection and I don't want to select the best features using evolutionary algorithms. I need to change my space to new reduced space.
Thank you János
Your link is incorrect. I can't go into it;
I don't know anything about your explanation. please refer me to a lecture,site or tutorial
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing