.\"Text automatically generated by txt2man
.TH hmm_train  "1" "" ""
.SH NAME
\fBhmm_train \fP- hidden markov model (hmm) training
.SH SYNOPSIS
.nf
.fam C
 \fBhmm_train\fP [\fB-h\fP] [\fB-v\fP] \fB-i\fP \fIstring\fP \fB-t\fP \fIstring\fP [\fB-b\fP] [\fB-g\fP \fIint\fP] [\fB-l\fP \fIstring\fP] [\fB-m\fP \fIstring\fP] [\fB-o\fP \fIstring\fP] [\fB-s\fP \fIint\fP] [\fB-n\fP \fIint\fP] [\fB-T\fP \fIdouble\fP] \fB-V\fP 
.fam T
.fi
.fam T
.fi
.SH DESCRIPTION


This program allows a Hidden Markov Model to be trained on labeled or
unlabeled data. It support three types of HMMs: discrete HMMs, Gaussian HMMs,
or GMM HMMs.
.PP
Either one input sequence can be specified (with \fB--input_file\fP), or, a file
containing files in which input sequences can be found (when \fB--input_file\fP and
\fB--batch\fP are used together). In addition, labels can be provided in the file
specified by \fB--label_file\fP, and if \fB--batch\fP is used, the file given to
\fB--label_file\fP should contain a list of files of labels corresponding to the
sequences in the file given to \fB--input_file\fP.
.PP
The HMM is trained with the Baum-Welch algorithm if no labels are provided. 
The tolerance of the Baum-Welch algorithm can be set with the \fB--tolerance\fP
option.
.PP
Optionally, a pre-created HMM model can be used as a guess for the transition
matrix and emission probabilities; this is specifiable with \fB--model_file\fP.
.SH REQUIRED OPTIONS 

.TP
.B
\fB--input_file\fP (\fB-i\fP) [\fIstring\fP]
File containing input observations. 
.TP
.B
\fB--type\fP (\fB-t\fP) [\fIstring\fP]
Type of HMM: discrete | gaussian | gmm.  
.SH OPTIONS 

.TP
.B
\fB--batch\fP (\fB-b\fP)
If true, input_file (and if passed, labels_file) are expected to contain a list of files to use as input observation sequences (and label sequences). 
.TP
.B
\fB--gaussians\fP (\fB-g\fP) [\fIint\fP]
Number of gaussians in each GMM (necessary when type is 'gmm'. Default value 0. 
.TP
.B
\fB--help\fP (\fB-h\fP)
Default help info. 
.TP
.B
\fB--info\fP [\fIstring\fP]
Get help on a specific module or option.  Default value ''. 
.TP
.B
\fB--labels_file\fP (\fB-l\fP) [\fIstring\fP]
Optional file of hidden states, used for labeled training. Default value ''. 
.TP
.B
\fB--model_file\fP (\fB-m\fP) [\fIstring\fP]
Pre-existing HMM model (optional). Default value ''. 
.TP
.B
\fB--output_file\fP (\fB-o\fP) [\fIstring\fP]
File to save trained HMM to (XML). Default value 'output_hmm.xml'. 
.TP
.B
\fB--seed\fP (\fB-s\fP) [\fIint\fP]
Random seed. If 0, 'std::time(NULL)' is used.  Default value 0. 
.TP
.B
\fB--states\fP (\fB-n\fP) [\fIint\fP]
Number of hidden states in HMM (necessary, unless model_file is specified. Default value 0. 
.TP
.B
\fB--tolerance\fP (\fB-T\fP) [\fIdouble\fP]
Tolerance of the Baum-Welch algorithm. Default value 1e-05. 
.TP
.B
\fB--verbose\fP (\fB-v\fP)
Display informational messages and the full list of parameters and timers at the end of execution. 
.TP
.B
\fB--version\fP (\fB-V\fP)
Display the version of mlpack.
.SH ADDITIONAL INFORMATION

For further information, including relevant papers, citations, and theory,
consult the documentation found at http://www.mlpack.org or included with your
distribution of MLPACK.