NLP techniques such as named entity recognition, sentiment analysis, topic modeling, text classification with Python to predict sentiment and rating of drug from user reviews.

In this project, the primary goal is to conduct analysis to determine how prescription drugs were rated by users based on the language of their reviews.

The analysis includes use of NLP and ML techniques in data cleaning and preprocessing, feature extraction and engineering, and data modeling.

The data contains patient reviews on specific drugs along with related conditions and a 10-star patient rating system reflecting overall patient satisfaction. The data was retrieved by crawling online pharmaceutical sites and later published in a study on sentiment analysis of drug experience over multiple facts including effectiveness and side effects of these drugs.

Text Classification to classify raw text reviews into negative or not negative reviews

Topic Modeling to discover latent semantics within the raw text reviews

Primary Metric - Area Under the Receiver Operating Characteristic Curve (AUC)

Seconday Metric - Accuracy, Precision, Recall, F1-Score

Bag Of Words (BoW)

Term Frequency-Inverse Document Frequency (TF-IDF)

Word2Vec Embeddings

Logistic Regression

Naive-Bayes

Random Forest

Gradient Boosting

Project_jnb_text_classification.ipynb - Contains python code used for text classification using the various ML classifiers mentioned above Project_jnb_topic_modeling.ipynb - Contains pythong code used for applying topic modeling method Project Report.pdf - Detailed report for the project

-- Aarif M Jahan -- Dec 12, 2021