Closed meghanabhange closed 4 years ago
Check out this pull request on
You'll be able to see Jupyter notebook diff and discuss changes. Powered by ReviewNB.
What's the acc number now? Did it change?
On Thu, 2 Jan 2020, 16:12 meghanabhange, notifications@github.com wrote:
@meghanabhange https://github.com/meghanabhange requested your review on: #19 https://github.com/NirantK/Hinglish/pull/19 Metrics.
— You are receiving this because your review was requested. Reply to this email directly, view it on GitHub https://github.com/NirantK/Hinglish/pull/19?email_source=notifications&email_token=AAYZUPMGTVYWAQG3NW5LDFDQ3XALBA5CNFSM4KCA3MTKYY3PNVWWK3TUL52HS4DFWZEXG43VMVCXMZLOORHG65DJMZUWGYLUNFXW5KTDN5WW2ZLOORPWSZGOVXSESJY#event-2917419303, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAYZUPKFDGMCR2KNMRSSS6LQ3XALBANCNFSM4KCA3MTA .
In general, the accuracy and F1 scores have improved.
Model | Before Processing | After Processing |
---|---|---|
NBSVM | Cross Validation Scores [0.59648356 0.60573225 0.60227845 0.60166627 0.58526608]Valid accuracy: 0.59853 Valid f1_score: 0.60262 | Cross Validation Scores [0.61221384 0.60546175 0.60388084 0.58896403 0.60085742]Valid accuracy: 0.62671 Valid f1_score: 0.63211 |
LR-TFIDF | Cross Validation Scores [0.5856514 0.60067245 0.58841432 0.59676788 0.58360738]Valid accuracy: 0.59485 Valid f1_score: 0.59924 | Cross Validation Scores [0.61372945 0.60484421 0.59632548 0.61229709 0.59192455]Valid accuracy: 0.61912 Valid f1_score: 0.62284 |
NB-SVM Regular Train Labels confident learnin | Cross Validation Scores [0.57748717 0.58891548 0.57213543 0.59675334 0.57944188] Valid accuracy: 0.59559 Valid f1_score: 0.59869 | Cross Validation Scores [0.58569709 0.58793323 0.575477 0.55926908 0.58868271]Valid accuracy: 0.60091 Valid f1_score: 0.60626 |
LR Regular Train Labels confident learning | Cross Validation Scores [0.59389899 0.61350341 0.60321008 0.6076949 0.58864606]Valid accuracy: 0.58897 Valid f1_score: 0.59264 | Cross Validation Scores [0.61478138 0.6073194 0.59535051 0.59933565 0.60500858]Valid accuracy: 0.62443 Valid f1_score: 0.62895 |