The words in a language not only help us to construct the sentences but also contain some other features, which we usually underestimate. Each word relates itself to the
remaining ones in some way. In our daily lives, we extensively use these relations in many areas, where question direction is also one of them.
In this work, it is investigated whether the relations between the words can be useful for question direction and an approach for question direction is presented.
Besides, a tool is devised in the way of this approach for a course given in Turkish.
The relations between the words are represented by a semantic network for nouns and verbs. By passing through the whole course material and using the relations
meronymy for only nouns / synonymy, antonymy, hypernymy, coordinated words for both nouns and verbs / entailment and causality for only verbs, the semantic
network, which is the backbone of the application, is constructed.
The end product of our research consists of three modules:
· / getting the question from the user and constructing the set of words that are related to the words that make up the question
· / scoring each course section by comparing the words of the question set and the words in the section
· / presenting the sections that may contain the answer
The sections that are evaluated are taken as the sections of the course for granted.
The chat logs that expand three years of the course were taken by permission and questions were extracted from them. They were used for testing purposes of the
constructed application.
| Identifer | oai:union.ndltd.org:METU/oai:etd.lib.metu.edu.tr:http://etd.lib.metu.edu.tr/upload/12605628/index.pdf |
| Date | 01 December 2004 |
| Creators | Sualp, Merter |
| Contributors | Turhan-yondem, Meltem |
| Publisher | METU |
| Source Sets | Middle East Technical Univ. |
| Language | English |
| Detected Language | English |
| Type | M.S. Thesis |
| Format | text/pdf |
| Rights | To liberate the content for public access |
Page generated in 0.002 seconds