I am using R to perform logistic regression on my data set. My data set has more than 50 variables.
The challenge is to write code in R that can assess the statistical validity of certain records and variables (e.g., p values >.05) and eliminate records and variables from the model based on parameters such as that.
Is there any already implemented way to do this? Any help or suggestion will be appreciated. Thank you.
Hi, Good Morning
For your study, My opinion is: don't take 50 variables. First you consult with your Principal Investigator (or) any doctors and find out the most related risk factor of a particular disease and take some other demographic variables for your study and do the logistic regression analysis.
In that, you have to find the significant variables (refer some previous journal papers related to your study) and eliminate or donot take or omit the insignificant variables from your study.
regards.
why would you want to eliminate any records? ok, you might need to deal with some outliers among the predictors, but even then you probably don't want to eliminate records.
A good way of doing variable selection in logistic regression is using LASSO or related penalization approaches. You can read about LASSO in http://statweb.stanford.edu/~tibs/lasso.html. You can fit logistic regression models using LASSO with the function glmnet of the glmnet package http://cran.r-project.org/web/packages/glmnet/index.html
- Badges
- Science topic
- Similar topics
- Engineering
- Computer Engineering