matrix - Linear or Nonlinear Regression of Multiple matrices with R? -

matrix - Linear or Nonlinear Regression of Multiple matrices with R? -

hello r friends , statistics thinkers !

hope question not stupid, wondering heal ignorance experiences or modeling knowledge. have many matrices of quantitative variables, allow me start 3 matrices of proportions. variabilities between samples, observe variables correlation between "contents m1", "variables m2" , "chemical m3" comparing these matrices. problem have 6 samples, tried sparse compositional model r package, not work because have little sample number n <10.

could please help me model selection , adequate r packages? should start iterative process trying little number of variable each matrice?

example of 3 matrices :

matrice 1

content1 content2 content3 content4 .... content150 sample1 sample2 sample3 sample4 sample5 sample6

matrice 2

v1 v2 v3 v4 ... v20 sample1 sample2 sample3 sample4 sample5 sample6

matrice 3

chemical1 chemical1 chemical1 .... chemical30 sample1 sample2 sample3 sample4 sample5 sample6

r matrix model regression sparse-matrix