Year: 2016
Type: dissertation
Abstract: Today, information overload has become a serious problem with the explosive growth of resources available on the Internet.Web users are commonly overwhelmed by huge amount of data and are faced with the challenge of finding the most relevant and reliable information in a timely manner.It is critical to use intelligent software systems to assist users in finding the right information from an abundance of Web pages.Furthermore, if researchers can quickly find out, from an article database, appropriate research papers and/or academic journals which they believe that they should take time to read just by using intelligent software, their time for searching will be saved.In order to solve the problem, this research proposes a model of multiple agents in which they gather data, exchange it together, and be ranked by the process of analyzing reliability using probabilistic reasoning.Belief and reliability among agents will be changed during the time they interact with each other.The process of ranking and classifying each agent is able to be trusted or not, therefore, are meaningful since they help us find out which one we should believe.From that result, important decisions can be made.Basically, the significance of this research can be divided into two main points.Firstly, this research provides us a method of classifying agents as reliable or untrustworthy.This classification method is a good technique for solving a wide range of problems.Secondly, a concrete implementation of an information filtering system will demonstrate the advantages of flexible application of probabilistic reasoning in analyzing the reliability of agents as well as the application of neural networks in extracting useful information.Using the system to download and filter webpages and/or articles, all what users see are articles that really interest him/her.This research will also have a variety of applications for solving other problems such as predicting game strategies, finding malfunction parts, analyzing detective stories, etc.This research aims to obtain two main goals.Firstly, the research aims to find a method of analyzing reliability of intelligent agents for ranking them when they exchange information together.In a simple view point, agents are intelligent autonomous computer programs (software).Autonomous software can be independent of human's control and has the ability to make its own decisions while intelligent software includes some qualities that the human mind has, such as the ability to understand language, to recognize picture, to solve problem and to learn.The second goal of this research is to construct an information filtering system that contains multiple intelligent agents, ranks them bases on their reliability and filters out useful data (find out reliable and trustworthy health information) gathered and exchanged by them using a combination of statistical probability and neural networks.i
Author: Tran Xuan, Hoang
Institution: Japan Advanced Institute of Science and Technology
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7047825605
