Using Semantic Models to Analyze Wikipedia Articles

The network architecture based on Web 2.0 is an example of the most common applications in use today. A famous example is Wikipedia, which is a free Internet encyclopedia that is created and maintained by general users. Like this site, users can change anything on it (subject to restrictions and further changes or reversions). Since the entry and maintenance of subject material can be edited by any Internet visitor, the information should not be considered to be complete or the final answer. Thus, this situation is easy to lead to the problem of information overload. The information overload always is caused by two main issues, synonyms (two terms are syntactically interchangeable expressions) and polysemy (a term has different meanings). In this paper, we adopt two semantic models, LSA (Latent Semantic Analysis) and PLSA (Probabilistic Latent Semantic Analysis) to solve these two issues. According to the results of simulation analysis, we conclude that using multiple semantic analysis techniques can give significant performance improvements.