Implemented Hotel Review Sentiment Pre-Built Model [resolves #257]

GlobalMaksimum / sadedegel

A General Purpose NLP library for Turkish

http://sadedegel.ai

MIT License

92 stars 15 forks source link

Closed onatyap closed 2 years ago

onatyap commented 3 years ago

A pre-built model for Hotel Review Sentiment was implemented in sadedegel/prebuilt/hotel_sentiment.py.
sadedegel/prebuilt/README.md was updated with explanation on loading and using model.
Benchmark results on the dataset were added to README.
Pre-built model with 86% accuracy on holdout test set was saved to sadedegel/prebuilt/model/ as hotel_sentiment_classification.joblib

dafajon commented 3 years ago

Can you please implement tests for the model before we merge.

dafajon commented 3 years ago

Also don't forget ~/.sadedegel_data/hotel_sentiment in hotel_sentiment dataset tests. Users without the dataset will experience fails.

husnusensoy commented 3 years ago

Can you explain results in SVM you compare ?

onatyap commented 3 years ago

I updated the comparison part in README using the values from the paper that first proposed HUMIR dataset. I think it is more clear now.

dafajon commented 2 years ago

Updated the branch with latest develop. All checks are passed. Model evaluation outputs the reported score of the metric. Merging and closing.