Quality oriented information retrieval and timeliness analysis on diabetes websites

---

Tezin Türü: Doktora

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Enformatik Enstitüsü, Bilişim Sistemleri Anabilim Dalı, Türkiye

Tezin Onay Tarihi: 2014

Öğrenci: RAHİME BELEN SAĞLAM

Danışman: TUĞBA TAŞKAYA TEMİZEL

Özet:

The foremost requirement of health information seekers is to retrieve high quality and up-to date information from web search engine results. Current techniques rely heavily on Web graph structure and they are domain independent solutions. However, in health domain, to ensure information quality, a search engine should return results that are not only relevant to submitted query but also in accordance with evidence based medical guidelines. The aim of this thesis is to propose an automated framework which is able to retrieve high quality and up-to date health information according to evidence based medicine. The contributions of the thesis are twofold. The first one is a method which is developed to differentiate high quality and unbiased health information content from low quality ones and makes use relevance feedback, information retrieval and opinion mining techniques. The second one is a method which is developed to automatically assess the timeliness of a health web site using evidence based clinical practice guidelines. The experiments are conducted on diabetes web sites and the results show that the first method achieves 76% accuracy in detecting high quality web sites and the second method accomplishes 77% in detecting the timeliness of web sites.