Analysis of enzyme interference factors in millet storage based on machine learning

Abstract

In this paper, we first investigate the peroxidase enzyme during millet storage, deeply analyze the characteristics of different types of grain bins during millet storage, and then summarize the peroxidase properties. Secondly, to extract the feature vector of the molecule, a descriptor was introduced, and on machine learning, SVM was used to construct a model of catalytic site MCD-MFEs and multiple catalytic sites SMAD-MFEs. Then, experimental materials were selected, experimental methods and measurement methods were determined, and an example analysis of machine learning-based enzymes during millet storage was performed, specifically from two aspects: model analysis and the study of peroxidase during millet storage. The results showed that the activity of millet peroxidase decreased by 92.2mg H2O2g−1, 90.4mg H2O2g−1, and 85.7mg H2O2g−1 for conventional, nitrogen-filled storage at 22°C. The activity of millet peroxidase decreased by 102.2mg H2O2g−, 98.8g H2O2g−, and 95.1mg H2O2g− The rate of reduction in peroxidase activity of millet stored in nitrogen-filled storage was not significantly different. This study was conducted to understand the enzyme change pattern during millet storage to provide a more intuitive and realistic reference for individual households to store grain.