Jetso providing system with web 2.0 tagging feature

Abstract

There are many web applications offering a place for users to share bargain or special discount information with others. In their system, we found some users placed the information in irrelevant category and only one category can be chosen to place the information. Because of these problems, users will miss some information if they just look at their interesting category. To solve the above problem, Web 2.0 tagging technology can be applied to enhance the information discovery. Tags are created by users in free from, so that tag quality is an important issue influencing the efficiency of information retrieval. In this project, we combine several research methodologies in our system in order to improve and maintain the tag quality so that the user can take advantage of tagging in information discovery. In this project, we have two study areas - information discovery and tag quality. The information discovery can be enhanced by relationship identification in tag searching and browsing. The tag quality can be improved and maintain by tag suggestion, rating widget, spell collection and tag education. The results of searching and browsing are bordered by finding out relevant tags via obtaining lexical relationship from WordNet database. To let user easily to discover the information, the tag cloud is created by length-normalized TF-IDF and bisecting K-means algorithms for selecting a higher coverage tag set and providing a new interface. Tag suggestion is generated by Tag Frequency Inverse Resource Frequency (TF-IRF) and Markov Clustering (MCL) algorithms. After keywords retrieving and filtering, the approach also can be used for no tag assigned information. Rating widget assists tag suggestion in determining which suggestion is better. An instant spelling correction service reduces tag divergence by lowering ratio of misspellings. Tag education equips users with tagging skill to unify the tagging style and make the tags useful. This project brings Web 2.0 Tagging technology into information retrieval system to increase the quality of information discovery and recovery.