When we recognize the existence or condition of something based on our five senses such as eyesight, hearing, touch, taste, and smell, we call the existence or condition a phenomenon.? I have a desire for understanding and elucidating a wide variety of phenomena in the real world.? My major preoccupation has always been what problems are essential and what philosophies and frameworks must be necessary to tackle the problems.? Among various phenomena, I am particularly interested in dynamic phenomena showing irregular and complex behaviours as time goes on.? I am also interested in “nonlinearity” where the change is not proportionate to input. There are still many phenomena which are not expressed in mathematical or physical formulae, as these basic principles are not understood well enough. To know the basic principles, although it will be preferable to study the organized mechanism (internal structural and operating principle) of the objects in detail in a natural state, it is not always possible and rather difficult. However, we typically have access to the behaviour of these systems. We can record the behaviours of the phenomena as numerical values by the observations and we can treat the collection of the values as data. We can investigate the characteristics (features and nature) using this measured data. Furthermore, as we can expect that the data is representative of the system, we can consider the essential mechanism based on the data characteristics and the situation when the phenomenon occurs. We can build a model which incorporates this idea and generate data using the model. By comparing between the simulation data and observational data, we can examine the appropriateness of the idea. In this way we can turn various thoughts over in our mind using data to elucidate phenomena. This form of data analysis is the focus of my research.? My main research is to unlock the underlying characteristics and mechanisms of the phenomena by data analysis without bias. My major approaches are following four.
I consider that spatiotemporal connection, time delay of communication, interaction of elements and nonlinearity are very important for these approaches. I consider that data analysis is useful, but also must be employed appropriately. Each approach of data analysis is based on a particular assumption about the underlying system. Models by data analysis are built statistically, and the condition of the simulations might be not identical to that of actual phenomena. Hence, it is difficult to consider that models and simulations obtained by data analysis are the same as the real things. On the other hand, when we look at the mechanism generating a phenomenon from the viewpoint of control and information, if the characteristics or behaviour of data obtained by data analysis are similar enough to those of the actual phenomenon, there is a possibility that the model and simulation capture the essence of the real mechanism. Data analysis is realistically possible one of the few effective methods when the basic principles of analysis objects are unknown and we cannot study them in detail. My research aims to elucidate phenomena using data analysis from this perspective.