Compatible Name showing under different group

[Sudx-old-gamer-new-coder]

hp_worksheet.csv.zip

A name is showing under different group using compatible names in Himachal AE.
Please see attached screenshot for details.

Also attaching the worksheet used.

[Sudx-old-gamer-new-coder]

The following name also didn't show up in the same group.

Cand | Constituency | Party | Year | pid

DES RAJ | BANIKHET | INC | 1972 | 497
DES RAJ | BANIKHET | INC | 1977 | 497
DES RAJ MAHAJAN | BANIKHET | INC | 1982 | 1244 <-( This row didn't show up in the group)

[hangal]

compat. names alg has 2 important params:

1. min token overlap -- if this # of tokens is the same in the 2 names, they are compatible. This is currently set by default to 3 (not 2). Will change it back soon, or make it configurable.

2. IGNORE_TOKEN_THRESHOLD = 200
   Any token that occurs more than this number of times in the dataset is ignored for the purposes of the above comparison. This is to get rid of noise due to extremely common names like Patel and Singh.
   In the log, look for a string like: Ignored tokens: SING RAM CHAND

[hangal]

Sudesh, pls check now and then close if fixed.
Compat. alg. now has 2 new params that can be controlled by the user.
Pls check the descriptions and suggest improvements to the text if not clear.
Default is 2, 200.

With these defaults, your examples work fine:

[Sudx-old-gamer-new-coder]

I had been playing around with the Compatible Name settings. But wasn't able to reduce the size of cluster which contains the following rows.
Could you suggest a setting that works for this?

BEERU RAM | | | | 2 | M | 23 | 0 | 9 | GEHARWIN | 3 | INC | 1998

BEERU RAM KISHORE | SC | | | 2 | M | 23 | 0 | 11 | GEHARWIN | 3 | INC | 2007
BEERU RAM KISHORE | | | | 1 | M | 23 | 0 | 8 | GEHARWIN | 3 | INC | 1993
BEERU RAM KISORE | | | | 2 | M | 23 | 0 | 7 | GEHARWIN | 3 | INC | 1990