The Jet time expression tagger identifies and normalizes time
expressions following the TIMEX2
standard. For example, "He
was born on July 4, 2008." would be tagged "He was born on <TIMEX2
VAL="2008-07-04">July 4, 2008</TIMEX2>." If the date is
not fully specified, it will be normalized based on a reference time, typically the
publication date of a news article. For example, in an article
published on July 4, 2008, the sentence "He was hired last October."
would be tagged "He was hired
<TIMEX2 VAL="2007-10">last October</TIMEX2>.
The tagTimex action uses
the time rules (in a file specified by property Timex.rule) to tag time
The reference time can be set in two ways. It can be given
explicitly (in ISO DateTime format) by property Timex.refTime.
Alternately, if the reference time is given by a field in the document,
action can be applied over that field of the document. The format
of the reference time in the document is specified by property Timex.refFormat, using the
syntax of Java SimpleDateFormat,
such as YYYY-MM-dd for a
year-month-day format. If a document collection includes
documents with different formats of reference time, a list of
SimpleDataFormats can be specified, separated by "|" (with no
blanks); for example, YYYY-MM-dd HH:mm:ss|YYYY-MM-dd
for a calendar date optionally followed by a time.