autoencoder_for_cancer

autoencoder to compress methylation data and cancer cell classifier

This project consists of 3 main parts

1. loading data with load_data folder: load methylation data from GDC database
2. preprocessor data with preprocessor.py: remove features with NA data, remove low-variant features, and separate data of normal/cancer into separated files.
3. classifier with model folder: compress data and classify if it is cancer or not autoencoder.py train an encoder model used for compression classifier.py train a classifier for compressed data predict.py uses the trained classifier and encoder to classify new data