Molecular_Solubility_Prediction_App

This app predicts the Solubility (LogS) values of molecules! Data obtained from the John S. Delaney. ESOL:  Estimating Aqueous Solubility Directly from Molecular Structure. J. Chem. Inf. Comput. Sci. 2004, 44, 3, 1000-1005.

Model Building Steps :

Step 1 : Importing Libraries and Dataset :

Important python libraries used in the project are Scikit-Learn , Pandas , Numpy and Streamlit . The Training dataset was obtained from the John S. Delaney. ESOL Dataset Link : https://raw.githubusercontent.com/dataprofessor/data/master/delaney_solubility_with_descriptors.csv

Step 2 : Data Preprocessing :

In this step the missing values in the data were checked.

Step 3 : Exploratory Data Analysis :

In this step Feature Scaling of the data was done.

Step 4 : Model Training :

A SVR model was trained using the medical data.

Step 5 : Model Prediction and Evaluation :

The model predictions were made .MSE and R-Squared values were used for evaluation of the model.

Website Link : https://molecular-solubility-streamlit.herokuapp.com/