[EMSA 040] Sentiment Classifier Benchmarking - Abhishekkalra88/Entity-Media-Sentiment-Analyzer GitHub Wiki
LOE:24hr(divided over the week)
As a developer: of the application, we wish to benchmark the performance of commonly available sentiment classifiers such as NLTK's Vader, Naive Bayes amongst others. Further, we wish to incorporate Anti Money Laundering (AML) , TF (Terrorist Financing) and Anti Bribery Corruption(ABC) keywords to broaden the definition of a negative sentiment.
Acceptance Criterion:
Input:
1)Available sentiment corpora such as NLTK's movie review repository for training the classifier. 2) Refine classifier performance by including Anti Money Laundering (AML) , TF (Terrorist Financing) and Anti Bribery Corruption(ABC) keywords. 3) Research the commonly available and used classifiers.
So that I: can increase the accuracy of the classifier to correctly classify the sentiment.
AC: Performance bench-marking by comparing precision, recall and F1 weighted scores across the classifiers.