此为数据挖掘课程大作业，主要功能是对已给定的描述生物分子三维结构的139351 个属性中，分类预测出该分子化合物对目标体的化学活性，从而达到帮助药物设计的目的。实验现有的数据包括：110个已标记活性标签的描述分子三维结构的139351维数据，634 个为标记活性标签的三维结构数据。希望根据这些提供的数据，采用适当的分类方法，提供出一个经过训练的分类器，能够尽可能准确的帮助未标记活性的分子化合物确定其活性。-This is a big job data mining courses, the main feature is the descr iption has been given three-dimensional structure of biological molecules of the 139,351 properties, classification to predict the molecular compound on the target body of chemical activity, so as to achieve the purpose of helping drug design. Experimental data available include: 110 has been marked activity of labeling a descr iption of three-dimensional structure of the 139,351-dimensional molecular data from 634 marked activity of labeling of three-dimensional structure of the data. Want to provide the basis of these data, the adoption of appropriate classification methods, a trained classifier, can be marked as accurately as possible without the help of active molecules to determine their activity.