Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 0 down vote favorite

Lets assume one has 2 datasets: with different number of rows (samples) and columns (features). Each of these 2 datasets have a column as a binary response variable. Lets say healthy or not. What sort of statistical methods can be used to help us for better feature selection results, or better performance in classification models, utilizing both datasets?

share|improve this question

1 Answer

up vote 0 down vote accepted

Put them together, call them a single dataset, add a column indicating which dataset they came from if that represents something meaningful you can measure out of sample (location?), proceed as normal with multiple imputation or another technique to deal with all the missing data this generates.

share|improve this answer
How to put together 2 datasets which differ in their attributes (columns)? E.g. the columns of the first dataset are x1,x2,..,x10 and the columns of the second are y1,y2,..,y20 where Xi and Yi are not the same variables. This resembles multi-modal data but the subjects (rows) are not the same. The only common column is the response variable. Regularized Canonical Correlation Analysis is the only statistics that seems to me a bit close to this problem. Do you mean a block diagonal structure where zeros complement the non-existent data? In this case what kind of statistics would be useful? – kmouts Mar 30 '12 at 6:01
all the rows of x1-x10 will be null in the second data set. All the rows of y1-y10 will be null in the second data set. sites.stat.psu.edu/~jls/mifaq.html – Patrick McCann Mar 30 '12 at 21:30
Thank you very much! – kmouts Mar 31 '12 at 19:00

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.