摘要
研究将手工定制和系统自动建模相结合的自适应神经网络建模方法,根据用户浏览网页的信息提取表示网页特征的关键词及权重,转换成代表用户兴趣特征的概念特征关键词,从而动态调整神经网络的参数,修正用户模型,使神经网络的输出能随用户的兴趣而改变。该模型能有效表示用户的长期兴趣,并随用户短期兴趣的改变调整神经网络模型。实验表明该方法能向用户推荐个性化网页,提高了推荐的精度。
It proposes a method of adapting neural-based user modeling through combining manual modeling and automatic modeling. According to the Web pages user browsed, the system extracts the key words and their weighs to represent the Web feature, and transforms them into the concept feature key words to represent user＇s interesting. The system adapts neural-based user modeling to recommend the personalized Web pages, and modifies parameters of the user modeling dynamically according to the users＂ interesting. Our experiments show that the user modeling method could express user＇s long-term interesting,and adjust along with the changes of user＇s short-term interesting.
出处
《电子测量技术》
2007年第4期165-168,共4页
Electronic Measurement Technology
关键词
神经网络
自适应
用户模型
个性化网页
neural network
adaptive
user modeling
automatic Web personalization
