ANALYSIS AND VISUALIZATION OF TIME-VARYING DATA USING THE CONCEPT OF 'ACTIVITY MODELING'

Scientific visualization, which transforms raw data into vivid 2D or 3D images and movies, has been recognized as an effective way to understand the large-scale datasets. However, most existing visualization methods do not scale well with growing data size. At present there are a number of techniques to analyze data belonging to a particular time-step, but not much research has been made into analysis of data with respect to correlation between time-steps. An attempt is made in this thesis to develop a system, extending the concept of Activity in DEVS, based on a simplified theory of finding the active regions in a cellular space of a spatio-temporal process. The process of analyzing and visualizing the time-varying data using the system is called Activity Modeling. Monitoring of activity would aid in analyzing the process with respect to it’s computationally efficiency, dynamically allocating resources to the then-active regions. Analysis of data using Activity Modeling gives a different perspective of the process under consideration and focuses only on the active regions in the cellular domain. An overall analysis of the process is presented, in the form of images and movies, of various results computed in the cellular and temporal domain The ‘Activity modeling’ system, apart from detecting active regions in the time-varying datasets, also computes statistical results and introduces a concept to predict a possible pattern based on the temporal correlation extracted from the data analyzed during the present time step.