阴性选择分类器原理与应用研究

发布时间：2019-01-01 12:23

【摘要】：本文主要研究利用人工免疫系统进行Web文本挖掘的方法。第1章首先对选题背景进行了介绍,然后介绍了免疫学的发展历程和主要研究内容。因为免疫学与人工免疫系统是密切相关的,对二者之间的内在联系进行了归纳。本文的研究内容就是建立在免疫学与人工免疫系统的关系基础上。第2章对自然免疫系统进行了较详细的介绍,主要包括免疫系统的基本组成、机制和原理。人工免疫系统所依据的免疫学原理主要包括免疫网络理论,克隆选择和阴性选择原理。人工免疫系统正是建立在免疫学理论和免疫系统机制基础上。第3章就集中介绍了人工免疫系统的二进制模型。许多是为了研究免疫系统机制而开发的。后来出现的模型逐渐转到工程领域。本章重点介绍了最早由Farmer提出的微分方程,基于基因库的模型和协同进化算法模型。协同进化算法模型在本文进行扩展,应用到Web文本挖掘。第4章主要对Web文本挖掘技术进行了详细讨论,Web文本挖掘技术是涉及多个技术领域的交叉领域。包括许多较为复杂的技术方法,从特征抽取到模型建立,以及模型评价方法等等。这一章与前三章尤其是第2和第3章结合起来,形成本文第5章所给出的Web文本分类模型理论与技术基础。第5章给出了基于协同进化算法的免疫阴性选择模型,并与传统方法进行了比较,给出了比较结果。表明人工免疫系统做为Web文本分类方法是可行的。虽然还有许多不足之处。
[Abstract]:This paper mainly studies the method of Web text mining using artificial immune system. The first chapter introduces the background of the topic, and then introduces the development of immunology and the main research contents. Because immunology and artificial immune system are closely related, the relationship between them is summarized. The research of this paper is based on the relationship between immunology and artificial immune system. Chapter 2 introduces the natural immune system in detail, including the basic composition, mechanism and principle of immune system. The immune principles of artificial immune system mainly include immune network theory, clone selection and negative selection. Artificial immune system is based on immunological theory and immune system mechanism. Chapter 3 focuses on the binary model of the artificial immune system. Many have been developed to study the mechanisms of the immune system. The models that emerged later gradually moved to the field of engineering. This chapter focuses on the differential equations proposed by Farmer, gene pool based model and coevolutionary algorithm model. The co-evolutionary algorithm model is extended in this paper and applied to Web text mining. In chapter 4, the Web text mining technology is discussed in detail. Web text mining technology is a cross domain involving many technical fields. It includes many complicated technical methods, from feature extraction to model building, and model evaluation methods and so on. This chapter is combined with the first three chapters, especially the second and third chapters, to form the theoretical and technical foundation of Web text classification model given in Chapter 5. In chapter 5, the immune negative selection model based on coevolutionary algorithm is presented, and the results are compared with the traditional method. The results show that the artificial immune system is feasible for Web text classification. Although there are many shortcomings.
【学位授予单位】：哈尔滨工程大学
【学位级别】：硕士
【学位授予年份】：2005
【分类号】：TP18;R392

【引证文献】