Subject Classifier is a classifier model trained on Multinomial Naive Bayes algorithm to predict subject given the query. Data was tested on other classification algorithms includeing Decision Tree, Random Forest, SVM. However, Multinomial NB was given preference as per the defined metric given the resources available and business problem to be solved. Data for subjects and related keywords was scraped from internet. Subject Classifier is used as an utility to Surfcourse for implementation of intelligent search mechanism.
Google Drive Links to Files:
classifier ( classifer.joblib.pkl) and Count Vectorizer ( cv.joblib.pkl ):
https://drive.google.com/open?id=1Lrf1vcOm0bYf7hNZzsuNW0Oty6sxebMD
Resources have been moved. Resources shall be provided over email. Please write to parth1102007@gmail.com