篇名 | Applications of Support Vector Machine Learning in Body Mass Index Prediction Based on the Gut Microbial Pattern |
---|---|
卷期 | 2:2 |
作者 | His-Chung Kung 、 Rouh-Mei Hu |
頁次 | 165-180 |
關鍵字 | Support vector machine 、 Microbiome 、 Obesity |
出刊日期 | 202008 |
Obesity is a risk factors for many chronic diseases, such as diabetes, cardiovascular disease, cancer, and virus infectious disease, such as Covid-19. Many animal experiments have demonstrated the correlation between the gut microbial pattern and obesity. Examination of the gut microbiota could be a novel approach for the prediction of obesity. To build a microbiome-based body mass index classifier, a set of gut microbiome data from 281subjects with a total of 986047 16S rRNA reads was used in this study. Wilcoxon signed-rank test demonstrated that there is no significant difference between the body mass index and microbial diversity and the Firmicutes/Bacteroidetes ratio. Correlation study identified that a weak positive correlation between body mass index and OTU 342467(Eubacterium biforme), r=0.201, and weak negative correlations with OTU 772282(Rikenellaceae), r=-0.258, OTU 191355 (Erysipelotrichaceae); and OTU 237444, r=-0.202 (Streptococcus), r= -0.256. A support vector machine body mass index classifier using a selected microbial subset as the classification feature was built. The prediction accuracy of this classifier was 82.59% for the training group and 80.70% for the testing group.