When using hurdle model, I got below info:
Error in optim(fn = countDist, gr = countGrad, par = c(start$count, if (dist == :
initial value in 'vmmin' is not finite
In addition: There were 50 or more warnings (use warnings() to see the first 50)
warnings()
Warning messages:
1: In sqrt(diag(object$vcov)) : NaNs produced
2: In sqrt(diag(object$vcov)) : NaNs produced
3: In sqrt(diag(object$vcov)) : NaNs produced
4: In sqrt(diag(object$vcov)) : NaNs produced
5: In sqrt(diag(object$vcov)) : NaNs produced
6: In sqrt(diag(object$vcov)) : NaNs produced
7: In sqrt(diag(object$vcov)) : NaNs produced
8: In sqrt(diag(object$vcov)) : NaNs produced
9: In sqrt(diag(object$vcov)) : NaNs produced
10: In sqrt(diag(object$vcov)) : NaNs produced
11: In sqrt(diag(object$vcov)) : NaNs produced
12: In sqrt(diag(object$vcov)) : NaNs produced
13: In sqrt(diag(object$vcov)) : NaNs produced
14: In sqrt(diag(object$vcov)) : NaNs produced
15: In sqrt(diag(object$vcov)) : NaNs produced
16: In sqrt(diag(object$vcov)) : NaNs produced
17: In pnbinom(0, size = theta["count"], mu = mu, lower.tail = FALSE, ... :
NaNs produced
18: In pnbinom(0, size = object$theta["count"], mu = mu, lower.tail = FALSE, ... :
NaNs produced
I used the same model for another project without such issue. The only difference between two projects is that current dataset includes ~1800 taxa while the previous one included ~200 taxa with prevalence of < 10-50% among the participants.
When using hurdle model, I got below info: Error in optim(fn = countDist, gr = countGrad, par = c(start$count, if (dist == : initial value in 'vmmin' is not finite In addition: There were 50 or more warnings (use warnings() to see the first 50)
warnings() Warning messages: 1: In sqrt(diag(object$vcov)) : NaNs produced 2: In sqrt(diag(object$vcov)) : NaNs produced 3: In sqrt(diag(object$vcov)) : NaNs produced 4: In sqrt(diag(object$vcov)) : NaNs produced 5: In sqrt(diag(object$vcov)) : NaNs produced 6: In sqrt(diag(object$vcov)) : NaNs produced 7: In sqrt(diag(object$vcov)) : NaNs produced 8: In sqrt(diag(object$vcov)) : NaNs produced 9: In sqrt(diag(object$vcov)) : NaNs produced 10: In sqrt(diag(object$vcov)) : NaNs produced 11: In sqrt(diag(object$vcov)) : NaNs produced 12: In sqrt(diag(object$vcov)) : NaNs produced 13: In sqrt(diag(object$vcov)) : NaNs produced 14: In sqrt(diag(object$vcov)) : NaNs produced 15: In sqrt(diag(object$vcov)) : NaNs produced 16: In sqrt(diag(object$vcov)) : NaNs produced 17: In pnbinom(0, size = theta["count"], mu = mu, lower.tail = FALSE, ... : NaNs produced 18: In pnbinom(0, size = object$theta["count"], mu = mu, lower.tail = FALSE, ... : NaNs produced
I used the same model for another project without such issue. The only difference between two projects is that current dataset includes ~1800 taxa while the previous one included ~200 taxa with prevalence of < 10-50% among the participants.
Any suggestions or comments are appreciated.