- Scikit-learn - ML library used
- Tensorflow Keras - ML library used
- Pytorch - ML library used
- ReactJS - Javascript framework used
- Pandas - Python Data Manipulation library used
- Seaborn - Data Visualization library used
This is the main file with all the preprocessing, EDA and Machine learning and Deep Learning Models.
- Installing libraries and Dependencies
- Importing dataset - UCI ML Drug Review Dataset
- Exploratory Data analysis
- Data preprocessing - Basic data information, cleaning up the data
- Dividing into test and train and transforming using Count Vectorise
- Applying Machine Learning models
- Applying Deep learning Models
- Applying Harvard Sentiment Dictionary Analysis
- Classifier Combination - Voting
This contains the emotional analysis done on the reviews using NRC Lexicon Library.
- It contains the same preprocessing as the above file.
- Post that NRC Lexicon library is explored.
- Reviews are passed to the library functions to get the emotion scores.
- Run the Pill_Recommendation.ipynb file first.
- The SVM code keeps crashing hence those cells should be avoided while running.
- LSTM takes about 1.5 hrs to complete running.
- The predictions from all the models are collected and stored in a .csv file.
- The final prediction scores calculated are also stored in a .csv file at the end.
- Run the Emotional_Analysis.ipynb file after that.
- It is a completely separate entity from the Pill_Recommendation.ipynb file. The results from both the files are used to predict data based on the reviews and rating as shown on the deployed website.
- It takes 6 hours to run.
| Name | Year | Branch |
|---|---|---|
| Harsh Agarwal | Sophomore | EE |
| Aditi Goyal | Sophomore | EE |
| Darshit Jain | Sophomore | EE |