Repository navigation
ChapmanMeanLength returns nearly the same value #5
Description
ChapmanMeanLength always returns 9.xx regardless of the inputs.
Here's the result when comparing "David" to the following associated strings:
David 9.223682
Dave 9.299208
Davdi 9.223682
Dadiv 9.223682
david 9.223682
Dovid 9.223682
Divad 9.223682
divaD 9.223682
Daves 9.223682
Maday 9.223682
xxxxx 9.223682
12345 9.223682
Dav id 9.148616
Da.v.id 9.07401
D-avid 9.148616
The first row should have a zero distance, but it returns a values which is the same to strings "xxxxx" and "12345".
I looked at the source code, and I don't see how this class could perform string comparisons.
Is this an incomplete class?
Metadata
Metadata
Assignees
Labels
No labels
Activity
StefH commentedon Aug 19, 2025
When I read the "description" from ChapmanMeanLength it seems that is not implemented as we would think.
I did ask ChatGPT for an implemenation, and I'll add this as
ChapmanMeanLengthTrueclass.David-Maisonave commentedon Aug 20, 2025
I get the sense that ChapmanMeanLength is not really a string matching algorithms. But it's true purpose is not clear to me.
FYI:
I'm using slightly modified version of the SimMetrics code in my own project.
SqliteFuzzyPlusExtension.
Most of the algorithms work great. ChapmanMeanLength is the only one I had issues with.