Data Mining Techniques to Understand Textual Data

Speaker Bio

Wubai Zhou is a Ph.D. candidate in Computer Science at Florida International University’s School of Computer and Information Sciences under the supervision of Dr. Tao Li and Dr. Shu-Ching Chen. Wubai received his B.Sc. degree in Computer Science and Technology in 2012 from Wuhan University, China and M.S. degree in 2015 from FIU. His research interests include system oriented data mining, system management, and machine learning.

Description

More than ever, information delivery online and storage heavily rely on text. Billions of texts are produced every day in the form of documents, news, logs, social network posts, etc. Text understanding contributes to many applications in the areas text summarization, search engine, recommendation systems, online advertising, conversational bot and so on. However understanding text for computers is never a trivial task, especially for noisy and ambiguous text such as logs, search queries.

Natural disasters such as hurricanes, earthquakes, and tsunamis cause inestimable physical destruction, loss of life and property around the world every year. In order to minimize the consequent loss of the disasters, an essential task in disaster management is to efficiently analyze, understand and summarize the disaster-related situation updates which usually can be gathered and extracted from a myriad of web documents. With respect to IT service management, one important component in service management, system monitoring, is capable of tracking the states of a system by collecting system statistics information such as the CPU utilization and the memory usage, and then report those states with a service ticket in the Incident, Problem, Change (IPC) system. The information accumulated in the ticket, which describes the symptoms of the corresponding problem, is used for problem resolving by system administrators. Intelligent understanding of tickets is critical for a high-quality service delivery in which relevant resolutions from historical tickets can be automatically retrieved to alleviate human efforts involved in IT service management. In this work, we study the unique properties of those text resources and propose different data mining techniques to achieve aforementioned goals.