Hi All, Does anyone could help me to find the right way to do the following: I fixed this linear model: lm(Obs~ A1 + A2 + A3 + A4 -1, data=data); where Ai are land cover classes areas. I have then the Ai coefficients, the structure of the residuals (it can be approximated by a normal distribution), the confidence intervals, etc. As a second step, given the results of the regression model, I want to find for each 'Obs' the better combination of Ai coefficients which minimize the residuals. My thoughts are to use the distribution of the residuals of the regression model and the confidence intervals to do that. Does anyone could tell me how I could manage that? Any help is welcomed ! I am right now trying to solve this using the lme function, but may be I am getting wrong! Thanks in advance, Jean-Paul