项目名称: 双参数间隔/不敏感支持向量机及其拓展
项目编号: No.61202156
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 计算机科学学科
项目作者: 彭新俊
作者单位: 上海师范大学
项目金额: 23万元
中文摘要: 支持向量机在自然科学、工程计算和社会科学等诸多学科中具有广泛的应用。本项目重点研究一类新型支持向量机- - 双参数间隔/不敏感支持向量机的理论分析、算法与应用。主要内容包括:双参数间隔/不敏感支持向量机的改进模型、几何意义与几何算法、理论泛化界、基于双参数间隔/不敏感支持向量机的特征选择算法以及其它方面的应用。本项目的研究成果将丰富和发展支持向量机的理论和算法,并为支持向量机在不同学科中的数据挖掘应用提供新的计算方法,具有十分重要的理论意义和应用价值。
中文关键词: 模式识别;支持向量机;参数间隔;改进模型;拓展应用
英文摘要: Support vector machines (SVMs) have been successfully applied in many disciplines, including the nature science, engineering calculation, and social science fields, etc. This project mainly researches a novel class of SVMs, called the twin parametric margin/insensitive support vector machines, including their the theoretical analysis, algorithms and applications. The main contents include: their improved models, the geometric interpretations and algorithms, the theoretical generalization bounds, and the novel feature selection method for real-world problems based on the twin parametric margin/insensitive support vector machines and some other applications. The research results of this project are of great theoretical significance and application value. They will not only fully enrich and develop the theory and algorithms of SVMs, but also provide some new methods for the applications of SVMs on the data mining of different disciplines.
英文关键词: Pattern recognition;Support vector machine;Parametric margin;Improved model;extension application