Identification of Gut Flora Based on Robust Support Vector Machine

Abstract

Gut flora parasitize the human gastrointestinal tract to maintain normal physiological functions, but can also lead to a variety of diseases and affect mental health. The relationship between gut flora and human health has attracted increasing attention and become a popular research hotspot at present. However, biological data are often characterized by large size and high dimensionality, leading to limited ability of traditional statistical-based methods to handle these data and make them difficult to analyze. In this paper, we use a support vector machine (SVM) model to perform data mining analysis on the Gut flora dataset. Specifically, we build a robust SVM model to compare the gut flora of obese and healthy people based on the public database of the American Gut Project, analyze the characteristics of the gut flora of obese people, and set up a machine learning model to predict the obesity status of people based on the gut flora to provide a theoretical basis for obesity intervention based on the gut flora. Different from the traditional support vector machine model, we use ramp loss to calculate the sample loss, avoid the influence of noise data, and get more accurate experimental results. The experimental results reveal the characteristics of the gut flora of obese people, apply machine learning to obesity prediction, and provide new research ideas and theoretical basis for precision diet and precision medicine.