ML and DL Approaches for Drug Review Classification Using a Composite Effectiveness–Satisfaction Score
Year: 2026
Type: article
Abstract: With the increasing availability of patient-generated content on platforms like WebMD, sentiment analysis has emerged as a powerful tool to understand user experiences with pharmaceuticals. This study presents a comparative evaluation of traditional machine learning (ML) and deep learning (DL) models for classifying drug reviews based on a novel composite metric-the Drug Effectiveness-Satisfaction Score (DESS)-calculated as the average of user-rated effectiveness and satisfaction. A diverse set of ML algorithms (Random Forest, SVM, XGBoost) and DL architectures (LSTM, CNN-LSTM, DistilBERT, and BERT) were tested using multiple feature extraction and embedding methods, including CountVectorizer, TF-IDF, GloVe, and transformer-based embeddings. Results show that BERT achieved the highest F1-score (87.22%), while XGBoost offered a strong balance between accuracy and computational efficiency. Despite robust results, performance was slightly below benchmarks reported in other domains, likely due to the complexity and ambiguity introduced by the DESS metric and the unstructured nature of real-world medical reviews. This work underscores the potential of combining ML and DL for healthcare sentiment analysis and highlights future opportunities in domain-specific fine-tuning, ensemble modeling, and explainable AI.
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7123733177