how to do feature selection?

I have made a term-document matrix from a large textual data set and i want to find out the feature from that.Can any body help me with a quick method/implementation to do it?I have tried to develop MI implementation but could not figure out how to do it. please help!!

Answer 1

[Image Analyst]

I have no idea. I don't know what a "term-document matrix" is. There are an infinite number of features and it totally depends on what you are looking at and what you want to find/measure/determine. Even is you narrowed down the field, such as to image processing, there are still an infinite number of features. I don't know what "MI" is, unless it's the US state of Michigan. Is it part of one of the toolboxes? You didn't list any toolboxes in the Products list below your question, so probably not, unless you forgot to. So basically I'm unable to help you at all for because of the lack of any actionable information.

[Tripti Malviya]

Basically i m working on text classification. Tmg is the toolbox which can be used and does everything but i want to do it manually. MI is one of the feature extraction method from the tdm(term-document matrix) of the data set. A term-document matrix is formed by,as the name suggests,terms as rows and the document in which it is occurring as the columns. If its clear to u and if anyone can help me!!