Discovery of Markers from Chromatogram Data by Machine Learning

Introduction

Techniques for identifying or classifying samples from chromatogram data by using machine learning algorithms have attracted considerable attention. In machine learning from chromatogram data, it is necessary to create a data table from peak intensity information, but the accuracy of the discriminant model may be reduced if a correlation exists between the sizes of the peaks or the sizes of the peaks are extremely different. For this reason, pretreatment of the data after creation of the data table of peak information is generally necessary in machine learning of chromatogram data. This article introduces an example of the workflow in discovery of discriminant markers from GC-MS scan data of food samples by using Python 3.7.

October 31, 2019 GMT