Prediction of Functional Groups Present in an Organic Compound

Project Description

To create a predictive model which can predict the composition and underlying structure (Functional Groups present in the compound) of a new organic compound from its FTIR data given being trained on structures of different compounds and their Infrared Spectroscopy.

Objectives

1. Understand the Chemical composition and internal structure of organic compound.
2. Use Fourier Transformed Infrared Spectroscopes of compounds to determine underlining chemistry.
3. Use Machine Learning techniques to perform FTIR data analysis.

Pipeline

1. Find suitable FTIR spectroscopes for various compounds along with their Internal Molecular Stucture.
2. Pre-process the FTIR Spectroscopes for elimination of noise as well as normalizing the data for featurizing along with creating a common wavelength IR Spectroscopy for all compounds.
3. Extract Functional groups in a compound from the Molecular structure files acquired for the compound.
4. Validate the data for Dimensionality Reducution to obtain features.
5. Build and Hypertune the parameters of a Nueral Network for prediction of Internal Structure of an organic compound.
6. Visualize the FTIR Spectroscopes and final model.

Note:

For the tech stack used in the project:

pip install requirements.txt

Hope this is helpful.

Thank You!