The First International Conference on Advances in System Simulation SIMUL 2009 September 20-25, 2009 - Porto, Portugal

Tutorials

A Functional Approach to Semantic Retrieval of Text Documents

Vitaly Klyuev, University of Aizu, Japan

Nowadays, the Internet is a major source of information for millions of people all round the world. For many users, the Internet has supplanted TV and newspapers. Searching is the most common task performed on the Web by the end user. There are many general purpose and topic-oriented search tools available on the net but finding appropriate information is still difficult and searching remains the frustrating task.

The key reasons for such a situation can be characterized as follows. Texts are considered as “bags of words” by search engines. Such a view resulted in two main outcomes:

• Search engines do not take into account semantics of text documents when perform indexing and searching. In other words, they simply loose semantics of documents.
• The language to express queries is artificial: a set of key words. Its semantics is very far from the semantics of any natural language. It is not adequate to express user information needs.

The functional approach was initially proposed for the Russian language. Recently it was applied to English. Its main feature is that it uses a functional representation of the sentence meaning. This model makes it possible to distinguish semantically related documents and their parts. Semantic analysis can be done quite accurate.

In this tutorial, we will review key methods applied to semantic retrieval, consider the functional approach in detail, and report results of our experiments.