Direct deduction of chemical class from NMR spectra

TL;DR

This paper introduces a CNN-based method for directly classifying chemical compounds from NMR spectra, aiming to automate and speed up the process of chemical identification without full structure elucidation.

Contribution

It demonstrates that deep learning, specifically CNNs, can effectively classify chemical classes directly from NMR data, outperforming traditional clustering and image registration methods.

Findings

CNN achieved high classification accuracy.

Traditional methods were less effective for this task.

Deep learning offers promising automation in cheminformatics.

Abstract

This paper presents a proof-of-concept method for classifying chemical compounds directly from NMR data without doing structure elucidation. This can help to reduce time in finding good structure candidates, as in most cases matching must be done by a human engineer, or at the very least a process for matching must be meaningfully interpreted by one. Therefore, for a long time automation in the area of NMR has been actively sought. The method identified as suitable for the classification is a convolutional neural network (CNN). Other methods, including clustering and image registration, have not been found suitable for the task in a comparative analysis. The result shows that deep learning can offer solutions to automation problems in cheminformatics.