Menu

#3 Wrong Feature Ids in RegexCountFeatures

open
nobody
None
5
2006-07-16
2006-07-16
No

The feature id in RegexCountFeatures may get wrong values.
<code>
patternOccurence[index] =
Math.min(maxSegmentLength,patternOccurence[index]);
f.strId.id = maxSegmentLength * (index+1) +
patternOccurence[index];
</code>

The maximum value of patternOccurence[index] is
maxSegmentLength, but in the feature id assignment, we
can specify value from 0 to maxSegmentLength-1.
The window of ids created for every index should
probably have size of maxSegmentLength+1(to accomobdate
0 to maxSegmentLength offset values)

-amit

Discussion


Log in to post a comment.