ML_Project

Enhancing Toxicological Testing through Machine Learning

Repo structure

• data_analysis.ipynb: all the preprocessing (find feature, extract them, first possible prepreocesses)
• baseline_model.ipynbs: simple classification and regression without smiles
• classification_with_smiles2.ipynb: preprocessing on smiles and classification on data with them
• ridge_polynomial_regression.ipynb: ridge (with polynomial expanded features) using data with smiles
• find_smiles.ipynb: show the R code to extract smiles from data
• matrix_fact (folder):
  • Factorization_preparation.ipynb: preprocess data (find score, build a sparse matrix) to be used for recommender system
  • Factorization_surprise.ipynb: factorization using surprise library
  • Factorization_Lab.ipynb: factorization using raw functions done in the homeworks (not good results, better consider surprise case).
• smiles_proc.py: functions to extrapolate data from smiles

All other files can be ignored.

Authors

M. Leone, G. Macchi, M. Vicentini, M. Baity-Jesi, K. Schirmer