节点文献

农业智能问答系统中的用户偏好研究

Research on User Profile in the Agricultural Intelligent Question Answering System

【作者】 赵全东

【导师】 王芳;

【作者基本信息】 河北农业大学 , 计算机应用技术, 2011, 硕士

【摘要】 随着Internet的发展,传统搜索引擎查准率低下的缺点不断的暴露出来,个性化服务便应运而生。用户偏好作为个性化系统的核心,逐渐受到了重视。本文对用户偏好的建模以及更新方法等进行了深入的学习研究,构建了基于用户偏好的农业智能问答系统的原型系统。该原型系统将用户的访问页面和Web日志作为挖掘分析对象,并不需要客户端用户过多的参与,可以由系统自动的分析推断出用户的兴趣来构建偏好模型并更新之。另外系统仿照人类遗忘规律对不同种类的偏好进行不同程度的遗忘。通过不断的更新用户模型,使之越来越贴近用户的真正意图。本文着重从以下几个方面进行了学习和研究:(1)全面研究了信息检索系统的关键技术,重点研究了构建基于本体的用户偏好模型的关键技术和核心问题。(2)对用户偏好模型进行了重点研究,探讨了用户偏好模型的表示方法,建模方法,更新方法等关键技术,并提出了一种新的用户模型更新算法。(3)在充分学习和研究了传统的农业相关网站的基础上,构建了农业智能问答系统的原型系统。(4)认真设计了测试数据,对文中的算法和用户偏好模型进行试验和分析,验证了其有效性。

【Abstract】 With the development of internet,the shortcomings of lower precision of traditional search engines have been revealed.Individuation service emerges as the times require.The user profile,as the core of personalized system, is growing recognition. This paper study and research on the modeling and updating method of user profile,and bulids the Agricultural Intelligent Question Answering System.The system excavates and analyzes the the accessed pages and Web logs,doesn’t ask the client users to participate in it too much. It can automatically analyzes and deduces user profile , subsequently builds and update the model of profile .Besides this system imitate human oblivious rules to forgets the different kinds of interests in varying degrees.Through regular updating the model, making it more and more to get close the user’sreal intentions.This article mainly studies and researches on from the following aspects:(1) The key technology of information retrieval system is comprehensively studied in this paper, the core problem and the key techniques how to construct Ontology-Based Preference Model is mainly studied(2) The user profile model is mainly studied in this paper,and the representation ,modeling method,updating method and so on is discussed,finally a new algorithm for updating the model is put forward.(3) Based on studying and research on the traditional agricultural websites, the Agricultural Intelligent Question Answering System is build。(4) Test data is designed in earnest the algorithm and the model is ,experimented and analysised, its effectiveness is verified.

节点文献中: