随着博客信息源成指数级的增长,在博客空间中的信息检索,知识发现等任务正面临着巨大的挑战.博客特有的格式为以博客为载体的数据挖掘任务带来不便.本文提出挑选最具代表性的m个博文构成的博文集对博客兴趣建模,挑选的标准保证博文集中博文的重要性和主题多样性,并根据这两个指标来构造博文评估函数,将其转换成实例选择优化问题求解.实验以博客分类为目标,表明通过本文方法预处理后的博客,能够降低时间复杂度,提高分类准确率. With the exponential growth of blog information sources, tasks such as information retrieval and knowledge discovery in the blog space are facing enormous challenges.The blog-specific format brings inconvenience to the data mining task with blog as the carrier. The most representative blog post consists of m blog posts. The selected criteria ensure the importance and theme diversity of the blog posts. According to these two indexes, we construct the blog post evaluation function and convert it into instances Select the optimization problem to solve.The experiment to blog as the goal, indicating that the method of preprocessing blog, can reduce the time complexity and improve classification accuracy.