| 基于SVM用户建模的核函数选择研究 |
| |
| 引用本文： | 方若宇,张琼,许慰玲,张虹.基于SVM用户建模的核函数选择研究[J].微计算机信息,2011(8):241-243. |
| |
| 作者姓名： | 方若宇 张琼 许慰玲 张虹 |
| |
| 作者单位： | 汕头大学工学院;汕头大学理学院;仲恺农业工程学院信息学院; |
| |
| 基金项目： | 汕头大学青年基金; 广东省轻工装备产学研示范基地开放课题的支持 |
| |
| 摘 要： | 用户建模是从用户偏好数据中建立用户偏好模型的过程,用户偏好数据具有系统运行初期的稀疏性和非线形的特点。支持向量机(Support Vector Machine,简称SVM)具有小样本学习、非线形处理的能力,是合适的用户建模工具。SVM的非线形处理能力主要依赖于核函数,采用不同的核函数进行建模对模型的预测效果有重大影响。本文重点研究核函数的选择对基于SVM建模方法的影响,从中选取了表现较优的小波核函数,构建性能突出的SVM进行用户建模。实验证明该建模方法可以有效地从小样本数据中学习用户偏好信息,建立反映用户真实偏好的用户模型。
|
| 关 键 词： | 用户建模 支持向量机 核函数 |
Selection of Kernel Function in Support Vector Machine Based User Modeling |
| |
| FANG Ruo-yu ZHANG Qiong XU Wei-ling ZHANG Hong.Selection of Kernel Function in Support Vector Machine Based User Modeling[J].Control & Automation,2011(8):241-243. |
| |
| Authors: | FANG Ruo-yu ZHANG Qiong XU Wei-ling ZHANG Hong |
| |
| Affiliation: | FANG Ruo-yu ZHANG Qiong XU Wei-ling ZHANG Hong(Engineering College,Shantou University,Shantou 515063,Guangdong,China)(Information College,Zhongkai University of Agriculture and Engineering,Guangzhou 510225,Guangdoing,China)(Scinece College,China) |
| |
| Abstract: | User modeling is the process of modeling user preference from selecting user preference data which is characteristic of sparseness and non-linear in system first operating period.Support vector machine,which is capable of small sample learning and non-linear processing,is a suitable way of user modeling.Non-linear processing of support vector machine mainly depends on kernel function.Therefore,selecting deferent kernel functions has significant impact on the effectiveness of user modeling.This paper focuses... |
| |
| Keywords: | User Modeling Support Vector Machine Kernel function |
| 本文献已被 CNKI 等数据库收录！ |
|
