Chemical Substance Classification using Long Short-Term Memory RecurrentudNeural Network

摘要

This paper proposed a chemical substance detectionudmethod using the Long Short-Term Memory of RecurrentudNeural Networks (LSTM-RNN). The chemical substance dataudwas collected using a mass spectrometer which is a time-seriesuddata. The classification accuracy using the LSTM-RNNudclassifier is 96.84%, which is higher than 75.07% of the ordinaryudfeed forward neural networks. The experimental results showudthat the LSTM-RNN can learn the properties of the chemicaludsubstance dataset and achieve a high detection accuracy.