Semantic Tag-based Profile Framework for Social Tagging Systems

Abstract

Social tags have recently become popular for use as keywords for searching and annotating Web resources. The main limitation of using social tags is the focus on the keyword and human interpretability rather than on semantics and computer interpretability. Hence, although social tags are useful for information sharing and organizing, they lack the semantic metadata needed for the reasoning and inference functions that facilitate a personalized Web resource search. This study addresses this issue by defining a Semantic Tag-based Profile Framework (STPF) for integrating ontology technologies into user profiles and Web resource annotations. The proposed STPF is applied in a novel algorithm for automatically generating user profiles by combining the traditional term frequency scheme and the semantic technology. A Semantic Tag-based User Profile Platform (STUPP) based on the proposed STPF and algorithm is also used as a broker for personalized Web resource searches made for individual needs. The experimental results show that the proposed approach has higher precision and recall than the traditional term frequency scheme.