Text-based Classification of Medical Conditions using Hybrid CNN-LSTM model

Using UCI ML repository's Drug Reviews dataset to predict condition based on medical reviews. Structure of directory:

• raw_data: contains raw train and test data from https://archive.ics.uci.edu/ml/datasets/Drug+Review+Dataset+%28Drugs.com%29
• cleaning.py: Script containing functions to pre-process texts
• preprocessing.ipynb : Notebook contain splitting and processing raw data
• modelling.pynb : Notebook using pre-trained word embeddings " GoogleNews-vectors-negative300.bin.gz." from https://code.google.com/archive/p/word2vec/ for predicting top10 conditions selected from raw data. Used sequential CNN-LSTM model.
• testData.csv: Processed data for testing
• trainData.csv: Processed data for training