Data Acquisition[Drug Review Dataset]

Data Preprocessing[Cleaning, Normalization, Tokenization, Feature Extraction]

Sentiment Extraction[Textblob / Polarity / Subjectivity]

Feature Engineering[Generating Linguistic & Structural Features]

Label Encoding[Conversion of Categorical Variables to Numerical]

Data Splitting[Train : Test = 80 : 20]

Model Training[LightGBM & RF with Recursive Feature Elimination]

Evaluation[Accuracy, Precision, Recall, F1-Score]

diagram 2