首页 | 本学科首页   官方微博 | 高级检索  
     

基于RBF网络的显示设计主观评价方法
引用本文:颜声远,于晓洋,张志俭,彭敏俊,杨明.基于RBF网络的显示设计主观评价方法[J].哈尔滨工程大学学报,2007,28(10):1150-1155.
作者姓名:颜声远  于晓洋  张志俭  彭敏俊  杨明
作者单位:1. 哈尔滨理工大学,仪器科学与技术博士后科研流动站,黑龙江,哈尔滨,150080;哈尔滨工程大学,机电工程学院,黑龙江,哈尔滨,150001
2. 哈尔滨理工大学,仪器科学与技术博士后科研流动站,黑龙江,哈尔滨,150080
3. 哈尔滨工程大学,核科学与技术学院,黑龙江,哈尔滨,150001
基金项目:国家自然科学基金 , 黑龙江省博士后科研启动基金
摘    要:主观评价指标权重的确定方法是人机界面评价中的一项关键技术,现有的人机界面主观评价方法难以摆脱个人和随机性因素的影响.提出的基于RBF网络的主观评价指标权重计算方法,利用人工神经网络的自组织、自学习与自适应特性对网络进行训练,使网络学习隐含在训练数据中人机界面主观评价指标的权重规律中,自适应调整主观评价指标的权重,克服了主观赋权法定权的随机性因素影响.文中建立了基于RBF网络的光柱表人机界面主观评价模型;研究了主观评价的样本数量、扩展系数、网络模型精度三者之间的相互关系.对不同训练样本数的光柱表人机界面主观评价RBF网络模型的分析表明,采用80个训练样本可以得到令人满意的评价精度.

关 键 词:人机界面  主观评价  RBF网络  显示设计  光柱表
文章编号:1006-7043(2007)10-1150-06
修稿时间:2006-10-26

Subjective evaluation of user interface design using an RBF network
YAN Sheng-yuan,YU Xiao-yang,ZHANG Zhi-jian,PENG Min-jun,YANG Ming.Subjective evaluation of user interface design using an RBF network[J].Journal of Harbin Engineering University,2007,28(10):1150-1155.
Authors:YAN Sheng-yuan  YU Xiao-yang  ZHANG Zhi-jian  PENG Min-jun  YANG Ming
Affiliation:1. Instrument Science and Technology Postdoctoral Workstation, Harbin Science and Technology University, Harbin 150080, China; 2. College of Mechanical and Electrical Engineering, Harbin Engineering University, Harbin 150001, China; 3. College of Nuclear Science and Technology, Harbin Engineering University, Harbin 150001, China
Abstract:A key component in subjective evaluation of a human-machine interface(HMI) is the decision on index weights.Unfortunately,it is very hard to avoid the influence of individual biases and other random factors in existing HMI evaluation methods.A radial basis function(RBF) network-based approach for calculating the index weights appropriate for subjective evaluation is proposed in this paper,based on the function's properties of self-organization,self-learning,self-adaptation,etc.An RBF was trained to calculate the appropriate index weights for subjective evaluation from training data.In this way the influence of individuals and random factors should be eliminated.An analysis of the accuracy of the subjective evaluation of the human-machine interface of a bar meter by using an RBF network was carried using different training samples.The number of samples needed,the expansion coefficiency,the accuracy of the RBF subjective evaluation network model and the relationships between them were researched.The results show that an RBF based subjective evaluation of a bar meter HMI is sufficiently accurate with 80 training samples.
Keywords:human-machine interface  subjective evaluation  RBF network  display design  bar meter
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号