Classification of 1D Signals Using Deep Neural Networks

Abstract

The absorbance spectrum technique is as old as the first alchemists. They sought to identify and understand their elixirs by looking at the color and opacity of solutions as different reagents were mixed, heated, and stirred. Today it remains the most widely used spectroscopic technique for studying liquids and gases due to its simplicity, accuracy, and ease of use. An absorbance spectrum can be used to identify substances or measure the concentration of a molecule in solution. In this work, PLSR, GBR, RF and CNN models are trained on absorbance spectrum data of different liquid solvents and concentration of a specific molecule is predicted. A large data set is collected to train and test the models. The proposed deep convolutional neural networks gave the best results.