| 摘要: |
| 目的 提出了一种基于多项眼动数据的拖拉机造型设计的评选模型。方法 使用Eye-link眼动仪采集了30名被试在对4款不同的拖拉机设计效果图的评价过程中的眼动数据，在眼动数据与主观评价值的多重共线性检验的基础上，分析BP神经网络，并建立了拖拉机造型设计的评估模型。结论 模型均方误差MSE=0.040，平均相对波动AVR=0.188，预测值和主观值的配对样本t检验P值远大于0.05。使用目标用户在体验过程中的眼动数据来评估拖拉机的造型设计，为拖拉机的造型设计评估提供了新的思路和方法。 |
| 关键词: 用户体验 眼动追踪 拖拉机造型 评选模型 主成分分析 人工神经网络 |
| DOI：10.19554/j.cnki.1001-3563.2018.08.033 |
| 分类号:TB472 |
| 基金项目:河南省软科学研究计划项目（182400410309）；国家自然科学基金资助项目（51375510） |
|
| The Selection Model of Tractor Appearance Design Based on Multiple Eye Movement Data |
|
ZHONG Qi1, PEI Xue-sheng1, GUO Gang2, XU Na2
|
|
1.Henan University of Science and Technology, Luoyang 471023, China;2.Chongqing University, Chongqing 400044, China
|
| Abstract: |
| A selection model is proposed based on multiple eye movement data. The red desktop eye track is applied to collect the data of 30 object users during the evaluation of 4 different tractors appearance design. Based on the correlation analysis between the subjective evaluation and eye movement data, using principal component analysis(PCA) and BP neural network establishes user experience evaluation model of tractor appearance design combined with mental and physical index. The MSE=0.040,AVR=0.188,the paired sample t test between predictive and subjective values(p>0.05). Eye movement date during experience is used to evaluate the tractor appearance design, which provides new thinking and method for evaluation of tractor appearance design. |
| Key words: user experience eye tracking the tractor appearance selection model principal component analysis artificial neural network |
