Olay Tespiti Problemi icin Akan Veri Isleme Platformu Kullanimi: Avantaj ve Kisitlarin Incelenmesi,

Sahin, Özlem Ceren
Tatbul, Nesime
Karagöz, Pınar
Social media and social networks are now used extensively for information and news sharing. They provide ability to convey news and information much faster than conventional media for sharing weather conditions, traffic accidents and other unexpected events and situations. For this reason, event detection from social media messages is an intensively studied research topic. In this work, we examined the usability and performance of a streaming data processing platform, the Apache Storm, for event detection problem. We used two techniques used in the literature for event detection. Both alternatives are coded on Apache Storm. Apache Cassandra is used as the intermediate data storage medium. The first method of event detection is based on tracking the frequency of the words in the messages, and it is based on the assumption that the words with a sudden increase in the frequency rate indicate an event. The second method is a clustering-based method and it uses a version of the hierarchical clustering algorithms adapted for the event detection problem. We elaborated on the use of the features provided by Apache Storm for both methods, and discussed the facilities and limitations provided. In addition, information on how to use the system created for simulation purposes for experimental analysis is also provided.
Citation Formats
Ö. C. Sahin, N. Tatbul, and P. Karagöz, “Olay Tespiti Problemi icin Akan Veri Isleme Platformu Kullanimi: Avantaj ve Kisitlarin Incelenmesi,” 2018, Accessed: 00, 2021. [Online]. Available: https://hdl.handle.net/11511/76292.