Introduction for “EMPeaks”, High-throughput Spectrum Analysis Assisted by Machine Learning

There are a wide variety of measurement techniques that produce spectra as output datasets. Existing data analysis software and some open-source macros are useful but not sufficient for non-experts to perform peak-fitting analysis and interpretation. Therefore, we have investigated analysis method using unsupervised machine learning for high-throughput and automated peak deconvolution analysis of spectra without prior knowledge of experimental techniques and material databases. In this paper, we will introduce the open-source package called “EMPeaks”, including development chronology, specific usage, current issues, and analysis examples.