As a researcher, I do research in Quantitative Social Research, Computational Linguistics and Artificial Intelligence. More specifically, I work on textual event information processing on social media, mainly on identifying relevant information, event information extraction, and time-to-event estimation.
Relevancer aims at finding a balance between complexity of a tweet data set and expert needs from this data. The clustering step identify samples of the information groups that are present in the data set. The expert’s opinion is used to annotate selected clusters. As a result, the annotated clusters of tweets can be used as they are for further analysis and to build a classifier that can label remaining tweets, which were not included in the annotated clusters, and any other tweet set in light of the view of the expert.
SInfEx developed to extract information from social media. It contains mainly grammars to detect traffic domain in Dutch.
Ali Hürriyetoglu, Florian Kunneman, Antal van den Bosch.
Proceedings of the 13th Dutch-Belgian Workshop on Information Retrieval
Publicly available microblogging platforms such as twitter.com, identi.ca, rstat.us, etc. and traditional news media taken into account. Differences between the organization of the information between different sources will be analyzed and bridged (e.g. to connect micro blogs and news releases to the same event) in order to benefit from the available data as much as possible.
My work is on language analysis for sentiment analysis in Turkish on mainly social media data. I start the analysis by determining morphologic and syntactic structure of a message. In the following step, I identify sentiment bearing linguistic structures and assign their sentiment value.
I worked as a consultant for a textual information extraction project as a freelancer. The task was to extracted border related events from news articles in Turkish.