关键词： 机器学习, 数据分析, 砂岩, 物理力学性质, 阈值温度

Abstract: In this paper, machine learning algorithm is adopted to determine the threshold temperature of physical and mechanical properties changes of sandstone post high temperature. Based on Python language, SVM and K-means algorithm are employed to classify and cluster sample data, determine the threshold temperature range, and verify the rationality and accuracy of dichotomies. The results show that 1) the threshold temperature range of the sandstone sample is 400 ℃ ~ 600 ℃; 2) there are significant differences on the physical and mechanical properties of the sandstone around the threshold temperature. The physical and mechanical properties of the rock samples are relatively scattered below the threshold temperature range, while the physical and mechanical properties of the rock samples are relatively concentrated above the threshold temperature range.

Key words: machine learning, data analysis, sandstone, physical and mechanical properties, threshold temperature