Knowledge Discovery and its Environmental Contribution from Multidimensional Time-Series Data

Physical phenomena and human behavior contain a variety of explicit and implicit laws. The methodology to discover this kind of laws by induction from observation and measurement data, called data mining, has become widely studied. Moreover, with regard to many phenomena in the world, there are time-series data that changes over time. My research interests include novel knowledge-discovery algorithms, such as self-organized configuration of visualization from multi-dimensional and large-scale data to capture the entire image of phenomena over time, and co-occurrence pattern mining algorithms to extract interaction between events. Moreover, I am also pursuing fundamental study on various evaluation indicators with an emphasis on user's interpretation towards the results of visualization, and meta-learning methods to improve the extraction accuracy by evolutionary method adapting to the target domain. On the other hand, as for specific applications, I have applied my methods into environmental contribution, such as visualization of damage process of fuel cells, and the discovery of the interactions between earthquakes or between the members of fuel cells.

• 2005 Specially Appointed Research Associate, The Institute of Scientific and Industrial Research, Osaka University
• 2007 Specially Appointed Assistant Professor, The Institute of Scientific and Industrial Research, Osaka University
• 2010 Ph.D. (Information Science) Osaka University
• 2010 Assistant Professor, The Institute of Scientific and Industrial Research, Osaka University
• 2015 Associate Professor, The Institute of Scientific and Industrial Research, Osaka University