Sensitivity Analysis of parameters with respect to output
I am new to machine learning and rapidminer. I am using rapidminer to do a sensitivity analysis on my inputs with respect to a output. I have roughly 2000 inputs and I would like to find out which of them can affect or impact the output the most. I have tired using the optimize selection operator to choose the most relevant features and the list turns out to be over 1000 features which is meaningless to my analysis. If possible, is there a way to find a list of most relevant features according to their weights? I know you can use weighting operators like weight by correlation or information gain, but these operators does not take into account of inter-correlation between the parameters. Is there a machine learning method to solve this issue and if so how can I do it in rapidminer?