Intelligent Web-History Based on a Hybrid Clustering Algorithm for Future-Internet Systems

Abstract

Internet grows larger year by year and makes users to be confronted with large quantities of data that they cannot fully comprehend. The ongoing transition from Web 2.0 to the Semantic Web makes the development of intelligent services with the ability to discern, classify and simplify web information of vital importance. In this paper we present a new model for web-history organizing in order to improve the user action over the Internet. Based on this model we proposed an application, delivered as a Google Chrome browser extension, which organizes the web-history into semantic clusters, providing the user with an easy-to-follow hierarchal structure. The paper covers the main algorithms in the field, offering a comprehensive critical analysis, such as document vectorization, relational clustering, fuzzy and genetic variations and the item-set-based approach. Our work consists of adapting these algorithms to support an ever-increasing set of input data. The result is a hybrid variation that rapidly offers an acceptable solution, which is optimized in time, a quality preserved during the extensive web explorations a user may undergo. A variety of test results is presented in the end, with under-stress behavior and a selection of user experience.