黄土物理力学性质的数据挖掘方法初探

就实际生产中的土工试验数据建立黄土的目标数据集和粘性土的对比数据集，离散数据实现属性的泛化处理；结合属性的物理力学含义，引入信息熵进行属性相关分析，确立了颗粒比重、压缩系数、压缩模量三个强相关的属性，依此挖掘黄土物理力学性质的概念对比描述，结果基本符合黄土的特点。

Data Mining of Physical and Technical Features of Loess

This paper sets up objective data set of loess and comparison data set of soil from the test data in practice, generalizes the attributes by means of scattering, introduces information entropy to analyze the correlation based on its physical and technical properties, chooses specific gravity, compressing coefficient and compressing module which have strong correlations. The results show that the concept analytical comparison of loess answers to the feature of loess according to the data mining.