r - glmmLasso error and warning -
i trying perform variable selection in generalized linear mixed model using glmmlasso, coming error , warning, can not resolve. dataset unbalanced, participants (ptno) having more samples others; no missing data. dependent variable binary, other variables (beside id variable ptno) continous. suspect generic happening, fail see , have not found solution in documentation or on web. code, adapted glmmlasso soccer example is:
glm8 <- glmmlasso(group~ndufv2_dctabl+gper1_dctabl+ esr1_dctabl+esr2_dctabl+klf12_dctabl+sp4_dctabl+sp1_dctabl+ pgam1_dctabl+ank3_dctabl+rasgrp1_dctabl+akt1_dctabl+nudt1_dctabl+ polg_dctabl+ adarb1_dctabl+ogg_dctabl+ pde4b_dctabl+ gsk3b_dctabl+ apoe_dctabl+ mapk6_dctabl, rnd = list(ptno=~1), family = poisson(link = log), data = stackdata, lambda=100, control = list(print.iter=true,start=c(1,rep(0,29)),q.start=0.7))
the error message displayed below. specficially, not believe there nas in dataset , unsure meaning of warning regarding factor variable.
iteration 1 error in grad.lasso[b.is.0] <- score.beta[b.is.0] - lambda.b * sign(score.beta[b.is.0]) : nas not allowed in subscripted assignments in addition: warning message: in ops.factor(y, mu) : ‘-’ not meaningful factors
an abbreviated dataset containing necessary variables available in r format , can downladed here. hope can guided bit how go on analysis. please let me know if there wrong dataset or cannot download it. appreciated.
just follow on @kristofersen comment above. indeed start
vector messes analysis up.
if run
glm8 <- glmmlasso(group~ndufv2_dctabl+gper1_dctabl+ esr1_dctabl+esr2_dctabl+klf12_dctabl+sp4_dctabl+sp1_dctabl+ pgam1_dctabl+ank3_dctabl+rasgrp1_dctabl+akt1_dctabl+nudt1_dctabl+ polg_dctabl+ adarb1_dctabl+ogg_dctabl+ pde4b_dctabl+ gsk3b_dctabl+ apoe_dctabl+ mapk6_dctabl, rnd = list(ptno=~1), family = binomial(), data = stackdata, lambda=100, control = list(print.iter=true))
then fine , dandy (i.e., converges , produces solution). have copied example poisson regression , need tweak code situation. have no idea whether output makes sense.
quick note: ran binomial distribution in code above since outcome binary. if makes sense estimate relative risks poisson may reasonable (and converges), need recode outcome 2 groups defined 1
, 2
, mess poisson regression.
in other words
stackdata$group <- stackdata$group-1
before run analysis.
Comments
Post a Comment