The rapid expansion of the Internet has resulted not only in the ever-growing amount of data stored therein, but also in the burgeoning complexity of the concepts and phenomena pertaining to that data. This issue has been vividly compared by the renowned statistician J.F. Friedman (Friedman, 1997) of Stanford University to the advances in human mobility from the period of walking afoot to the era of jet travel. These essential changes in data have brought about new challenges in the discovery of new data mining methods, especially the treatment of these data that increasingly involves complex processes that elude classic modeling paradigms. “Hot” datasets like biomedical, financial or net user behavior data are just a few examples. Mining such temporal or stream data is a focal point in the agenda of many research centers and companies worldwide (see, e.g., (Roddick et al., 2001; Aggarwal, 2007)). In the data mining community, there is a rapidly growing interest in developing methods for process mining, e.g., for discovery of structures of temporal processes from observed sample data. Research on process mining (e.g., (Unnikrishnan et al., 2006; de Medeiros et al., 2007; Wu, 2007; Borrett et al., 2007)) have been undertaken by many renowned centers worldwide1. This research is also related to functional data analysis (see, e.g., (Ramsay & Silverman, 2002)), cognitive networks (see, e.g., (Papageorgiou & Stylios, 2008)), and dynamical system modeling, e.g., in biology (see, e.g., (Feng et al., 2007)). We outline an approach to the discovery of processes from data and domain knowledge. The proposed approach to discovery of process models is based on rough-granular computing. In particular, we discuss how changes along trajectories of such processes can be discovered from sample data and domain knowledge.
Domain knowledge integration in data mining using decision tables: case studies in churn prediction
