Solubility_model_development

Building regression model based on the CRISP-DM methodology. Comparing different regression algorithms and choosing the best one.

Selected algorithms (from scikit-learn, TensorFlow, Keras):

• linear
• Lasso
• Ridge
• elastic nets
• decision tree
• random forest
• gradient boost
• SVR
• MLP
• DNN

I used hyperparameter tuning and k-fold cross validation.