Credibility analysis on tweets for news and discussion programs by using a hybrid credibility analysis method

---

Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü, Türkiye

Tezin Onay Tarihi: 2016

Öğrenci: ALİ FATİH GÜNDÜZ

Danışman: PINAR KARAGÖZ

Özet:

In this work, we have studied credibility analysis of microblogging about news and discussion programs broadcast on television. We collected our data from one of the most important microblogging services, Twitter. Our credibility definition is based on three dimensions: being free from slang words, free from spamming purposes and newsworthy or important. We developed a hybrid model of supervised learning ap- proach and graph based hub and authority score transferring approach. Firstly apply- ing feature based classification on the collected data set and obtaining initial results, we tried to improve classification performance by graph based part of our study. Our graph based improvement approach is proposed to uncover the credibility relevance between microblogging messages and writers of those messages. We focused on message-message, message-writer and writer-writer connections in this graph. The performance of the proposed method is analyzed through a set of experiments. The final credibility score of a message is deduced based on each three dimension results at the end.