Data Reduction Algorithms Inquiry.

1. What would be good algorithms for identifying contributible parameters to an output? 2. What would be good algorithms for reducing data points while still retaining reasonable quality?

Thanks very much for your help in advance.

[dong huang]

Ex for 1.:I have a data log which contains 100s of parameters. Among these I am interested in 1 parameter and I want to know which parameters are contributible to this parameter. Probably PCA mnight qualify for this but I would like to know if there are more algorithm than this.

Ex for 2.: I have a data log which contains millions of data points. Among these data points, some are just repeats, some are similar, and some good information. Data this size is too large to be manageable, so it needs to be reduced. Roughly, K-means can help but I was wondering if there are better algorithm than it.

[KSSV]

You can remove repeated parameters using unique. And to select dominant features use pca.