An Intelligent Agent System for Topic Tracking and Classification

In this work, I proposed a model that can classify search engine results based on similarity of their topics. The Structural System Analysis and Design Methodology was used to analyze our findings. Vector Space Model was used to extract the set of components in a document, and probabilistic word model was used to find out cluster label for document classification. Clustering technique is used to group the documents into groups of similar topics for specific knowledge. The essence of applying these techniques is to build an intelligent information retrieval system which cluster internet documents into similar topic using an unsupervised machine learning techniques to reduce the percentage of irrelevant documents that are retrieved and presented to the user. The result shows that clustering specific concept helps users to visualize search engine results in a manner that allows the user to choose relevant pages effectively and to discover knowledge on the web in a way similar to a traditional book, to assist learning and reduce information overload.