quickSentiment 0.3.7

Bug Fixes

• TF-IDF Weighting Consistency: Fixed a mathematical mismatch in BOW_train and BOW_test. Previously, BOW_train utilized quanteda::dfm_tfidf (which defaults to standard inverse document frequency), while BOW_test applied a manual smoothed IDF formula (log((N+1)/(df+1)) + 1). The pipeline now correctly and consistently applies the smoothed scikit-learn style IDF weights across both training and testing datasets using quanteda::dfm_weight().

Minor Improvements

• Dependency Reduction: Removed the @importFrom dependency on the Matrix package, as the updated IDF weighting logic now operates entirely natively within the quanteda ecosystem.
• Fixed a minor missing comma syntax error in the switch statement of BOW_test.