技术基于机器学习的文本自动归类系统设计与实现.pdf

摘要 本文设计并实现了一个基于机器学习算法的文本自动分类系统，以新浪新闻内 容作为初始分类数据，使用支持向量机为基分类器，构造出一个能够分类文本信息 的分类系统。 本文所设计的分类器是通过有监督学习而成的分类器，本文对分类系统的原理 做了重点介绍。本文所设计的分类器基于支持向量机算法（SVM），并详细分析 SVM 算法的原理以及在文本分类领域中的应用。 本文介绍了文本的自动分类技术，并对预处理、中文分词、文本特征提取，模 型的训练和分类四个模块做出介绍。通过探究分类器自身参数和 SVM 核函数的选 择对分类效果的影响，并在此基础上进行改进，最终设计出了较好的中文文本分类 模型。 本文最后介绍了文本分类技术在目前发展的前景，并对未来文本挖掘技术提出 了期望和展望。 关键词：SVM；机器学习；文本分类；核函数 I Abstract In this paper an automatic text classification system based on machine learning algorithm is designed and implemented. Sina news content is taken as the initial classification data and support vector machine is used as the basic classifier to construct a classification system capable of classifying text information. The classifier designed in this paper is one based on supervised learning. The classifier designed in this paper is based on support vector machine algorithm (SVM), and the principle of SVM algorithm and its application in text classification are analyzed in detail. This paper introduces the automatic text classification technology and four modules: preprocessing, Chinese word segmentation, text feature extraction, model training and classification. By exploring the influence of the classifiers own parameters and the selection of SVM kernel functions on the classification effect, and improving on this basis, a better Chinese text classification model is finally designed. At last, this paper introduces the prospect of text classification technology at present, and puts forward the expectation and prospect of text mining technology in the future. Key words: SVM; Machine Learning; Text classification; Kernel function