论文部分内容阅读
为降低主观因素干扰,更加客观地评估产品造型的用户体验性,基于眼动数据,提出一种应用遗传算法优化BP神经网络的用户体验预测模型。将Tobii 120型眼动仪作为试验仪器,以Likert五级主观量表法作为辅助研究方法,采集了40位用户对12个工程车设计方案评价过程中的眼动数据,通过遗传算法对初始值编码,优化BP神经网络,建立眼动数据与主观评价相结合的综合评价模型。经广义线性回归模型筛选,确定了以注视时间为代表的10个眼动参数作为神经网络构建参数,随机选取35组眼动数据进行预测,结果显示该神经网络能有效预测产品造型设计用户主观体验得分,预测相对误差约5%。基于遗传算法的优化BP神经网络模型对使用眼动数据预测用户体验主观评价效果显著,可为今后产品造型的用户体验评价提供参考。
In order to reduce the interference of subjective factors and evaluate the user experience of product modeling more objectively, this paper proposes a user experience prediction model based on ophthalmic data to optimize BP neural network using genetic algorithm. Taking Tobii 120 eye tracker as a test instrument and Likert five-level subjective scale as a supplementary research method, 40 eye movement data collected by 40 users during the evaluation of 12 engineering vehicles were collected. The initial values Coding, optimizing BP neural network, establishing a comprehensive evaluation model of eye movement data and subjective evaluation. By generalized linear regression model screening, the 10 parameters of eye movement represented by gaze time were determined as the parameters of neural network construction, and 35 sets of eye movement data were selected randomly for prediction. The results showed that this neural network can effectively predict the users subjective experience of product design Score, predict the relative error of about 5%. The optimized BP neural network model based on genetic algorithm has significant effect on subjective evaluation of user experience by using eye movement data, which can provide reference for user experience evaluation of product modeling in the future.