Entering edit mode
8.2 years ago
toyan.j.p
▴
30
Hi,
I tried using glmnet package of R for obtaining coefficients. Though I have done linear regression of the same data set(49 SNPs, 50 samples) with a quantitative trait, I am a novice in this. Have a few queries on these and thanks in advance for any help
- What exactly is the lambda used in this what is its significance? Though I have got some interesting pics, I am not able to interpret it.
- How would you know which of the SNPs are included in the model? Is there something equivalent to R squared which gives the % of variance in QTL explained by each SNP in the model
- What is the simple interpretation of Coefficients Vs L1 Norm plot(and what is L1 Norm)? What is the top axis? How to give labels of column headers (in this case SNP ids) to each line (presently it gives variable number-I hope)?
- How to interpret Mean -Squared Error Vs log(lambda) plot used in cross validation?
- How to interpret coefficients Vs fraction deviation explained plot
- And what are the '.' seen when you run coef(cvfit, s = "lambda.min")?
