Text Categorization

mlt

Mlt tool is designed to tokenize fields from MEDLINE citations. Title, Abstract, and MH (starred MHs and SHs only) fields, and combinations of these are routinely tokenized and extracted from a MEDLINE citation. Other fields may be specified for tokenization as well.

Set Up

Follow the installation instructions to install text categorization tools and run the mlt program. Check on the following items only if you don't use the provided script to install Text Categorization tools.

CLASSPATH:
1. include the Text Categorization tools distribution jar file, ${TC_DIR}/lib/tc2011dist.jar, in your CLASSPATH.
2. include the TC top directory in your CLASSPATH.
Configuration File: assign the full path of the top directory of tc2011 to a variable named ROOT_DIR in the configuration file, data/Config/tc.properties.

Test Run

Run java program

Enter the command:


> mlt -h

Synopsis:
  mlt [options]

Description:
  mlt is a program to tokenize MEDLINE citations by specifying field tags

Options:
  -ci       Show configuration information
  -h        Print program help information (this is it)
  -i:STR    Specify input file (must specify)
  -pmid     Preserve PMID in the first field
  -s        Sort output by PMID
  -o:STR    Specify output file (must specify)
  -t:STR    Specify MEDLINE field tag:TI|AB|TIAB|MHs|TIABMHs|ALL (must specify)
  -v        Print the current version of mlt
  -x:STR    Specify an alternate configuration file

where:

mlt: Mlt script to run Mlt Java class
-h: set Mlt system option to display help information

Input

Three inputs must be specified when run mlt:

input File: MEDLINE citations

output File: where the results go

field tag:

TI: title
AB: abstract
TIAB: title and abstract
MHs: starred MeSHs
TIABMHs: title, abstract, and starred MeSHs
ALL: all fields (the output should be the same as the input)
Any legal tag in MEDLINE citations

Output

Each field will be sent to output and separated by line separator.

Mlt Options

Please refer to design document