摘要
个性化信息检索与获取是目前理论与应用研究的一个热点.其关键在于如何体现用户个性化以及有效缓解用户疲劳、加快算法的收敛.本文以图形检索为应用背景,提出了基于交互式遗传算法的用户Agent模型.该软件Agent针对现有研究的不足,将用户个性化信息获取与个性化检索集成在一起,两者相辅相成.在获取用户个性化信息时,我们设计了一种结合归纳和统计的用户情感计算机制,通过对前几代操作的结果进行归纳、计算,得到用户的特异性偏好;在利用用户情感偏好实现检索时,我们设计了利用个体偏好的引导进化方法来指导交互式遗传算法的选择、变异等操作.实验验证该模型在人脸图形检索中确实达到了体现用户个性化,有效缓解用户疲劳的目的.
The individual information retrieval is a hotspot of theoretic and applied research. Its key point is how to
embody the user's individuation, alleviate the user's fatigue, and quicken the arithemetic's convergence. This
paper takes graphic information retrieval as application background and provides User's Agent Model based on
IGA. Aimming at the shortage of existing research, this software Agent combines the user's individual ioformation
acquiring with individual retrieval. We design an affectional interest computing mechanism using induction and
statistic in early operating results to acquire the user's individual information, and we design a method which
conducts evolution using individual affectional interest to direct the selector and crossover. Experiments on
human facial graphic information retrieval validata its effectiveness in embodying the user's individuation and
alleviate the user's fatigue.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2004年第2期244-249,共6页
Pattern Recognition and Artificial Intelligence
基金
国家自然科学基金(No.60204009)
关键词
交互式遗传算法
用户Agent
个性化
情感
图形检索
Interactive Genetic Algorithm
User's Agent
Individuation
Affectional Interest
Graphic Retrieval
