gui11aume / hmmt Goto Github PK

Hi,
First, thank you for sharing this useful package.

I encounter an underflow error where scr/viterbi.f vit() complained that log(initialProb) was infinite.

In R/Viterbi.R I added

initialProb <- log(initialProb)
initialProb[is.infinite(initialProb)] <- -325

As it is done for emissionProb. Of course, I then gave initialProb in the argument of the fortran vit function.

I had never encountered this error before.

Cheers.
Jeremie

Hello!

I tried to specify the transition matrix using Q with my own state transition probabilities but once I run BaumWelchT, the state transition matrix has different values. Is this normal?

I have trouble installing the package, both when using devtools install_github, and when doing it by first cloning followed by R CMD INSTALL.

I get these error messages:

• installing source package ‘HMMt’ ...
  fatal: Not a git repository (or any parent up to mount point /tmp)
  Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).
  \nGit repository is dirty.
  Commit or clean changes before install.\n
  ERROR: configuration failed for package ‘HMMt’
• removing ‘/home/hub_kind/sperricone/R/x86_64-unknown-linux-gnu-library/3.1/HMMt’
  Error: Command failed (1)

It appears to be caused by the configure script.