Machine Learning Prediction of Empirical Polarity Using SMILES Encoding of Organic Solvents

Polarity_preprocessing.ipynb

Code consisting of all the feature preprocessing.

Polarity_214+1.ipynb

Final code based on 214 SMILES based descriptors and 1 categorical descriptors.

Polarity_7+1.ipynb

Final code based on 7 SMILES based descriptors and 1 categorical descriptors.

nn.h5

Saved final ANN model which can be loaded for prediction purposes.

polarity_raw.csv

Raw dataset consisting of 431 data points, 1613 numerical descriptors extracted from Smiles notations,

and SMILES notations of all the 431 molecules.

train.xlsx

Training set consisting of 378 data points, 214 smiles based descriptors, Name of the molecules,

solvent type, Et(30) values and SMILES.

test.xlsx

Test set consisting of 53 data points, 214 smiles based descriptors, Name of the molecules,

solvent type, Et(30) values and SMILES.