中文摘要：

为了解决网上用户的识别问题,研究了基于眼动的隐马尔可夫模型（HMM）的用户识别方法。使用眼动装置获取用户网上行为的眼动数据,并提取显著性眼动特征。使用隐马尔可夫模型分别对不同类型用户建立用户模型,并将用户数据输入模型。利用最大概率原则输出用户类型,并使用优化算法-遗传算法（GA）对HMM进行参数优化,提高了识别准确率。实验结果表明,通过该方法识别网上用户类型是可行的。该研究进一步表明,根据用户的网上行为特点,优化网页结构,能够满足不同用户的个性化需求,还可以对用户的个体行为进行独立挖掘,提高人机交互水平。

英文摘要：

To solve the problem of Web user identification,the method using Hidden Markov Model（ HMM） is explored dealing with the data of eye movement.The eye movement data of users＇ online behavior are acquired and the significant features of eye movement are extracted.Then The models of different types of users are established using HMM,inputting the data of users into model. The maximum probability principle is applied to output the user type and the Genetic Algorithm（ GA） is used to optimize the parameter of HMM,improving the accuracy of identification. The experimental results indicate that the Web users can be recognized effectively by the HMMmethod.In addition,according to the accurate user identification based on characteristics of user＇s online behavior,the structure of web pages can be optimized to meet the need of different users.It also can make the user＇s individual behavior independently and improve the level of human-computer interaction.