I'm studying allometric relationship between traits of insects. The pattern of the relationship from scatter plot is not clear. So, I would like to find the best fit model using regression analysis by comparing the data to various statistical models. I used lmodel2 and smtr packages for R, but they seemed work only with linear regression. What should I do if I would like to find equations from more complex models (such as polynomial and logistic) using reduced major axis method in R ?
Thank you in advance
I would suggest variable selection by adaptive lasso for least squares regression and logistic regression unless you have categorical predictors. In that case, I would recommend group lasso. See the attached paper for references. Best wishes.
Hi. Supose we have N values of only 2 variables k1 and k2, with medias U1 and U2. Divide each of N values K1i/U1 and do the same for K2i/U2. Result is two samples expressed in medias of each distribution. Then apply the non-parametric method you prefer, in order to correlate them in medias. The expression that links K1(medias, time fraction) = Constant* Y(k2 medias, time fraction). If you ignore the constant or asume it equal to one, graph shapes become more clear, but less precise in terms of physical units used. Try first with samples of few N
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling