Discovery of Process Models from Data and Domain Knowledge: A Rough-Granular Approach

Discovery of Process Models from Data and Domain Knowledge: A Rough-Granular Approach

Hung Son Nguyen (Warsaw University, Poland), Andrzej Jankowski (Institute of Decision Processes Support and AdgaM Solutions Sp. z o.o., Poland), James F. Peters (University of Manitoba, Canada), Andrzej Skowron (Warsaw University, Poland), Jaroslaw Stepaniuk (Bialystok University of Technology, Poland) and Marcin Szczuka (Warsaw University, Poland)
DOI: 10.4018/978-1-60566-324-1.ch002
OnDemand PDF Download:
List Price: $37.50


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.
Chapter Preview

Introduction: Wisdom Technology (Wistech)

In this section, we discuss a research direction for discovery of process models from sample data and domain knowledge within the Wisdom technology (wistech) system presented recently in (Jankowski & Skowron, 2007; Jankowski & Skowron, 2008a).

Wisdom commonly means rightly judging based on available knowledge and interactions. This common notion can be refined. By wisdom, we understand an adaptive ability to make judgments correctly (in particular, correct decisions) to a satisfactory degree, having in mind real-life constraints. The intuitive nature of wisdom understood in this way can be metaphorically expressed by the so-called wisdom equation as shown in (1).

wisdom = adaptive judgment + knowledge + interaction.(1)

Wisdom can be treated as a special type of knowledge processing. To explain the specificity of this type of knowledge processing, let us assume that a control system of a given agent Ag consists of a society of agent control components interacting with the other agent Ag components and with the agent Ag environments. Moreover, there are special agent components called as the agent coordination control components which are responsible for the coordination of control components. Any agent coordination control component mainly searches for answers for the following question: What to do next? or, more precisely: Which of the agent Ag control components should be activated now? Of course, any agent control component has to process some kind of knowledge representation. In the context of agent perception, the agent Ag itself (by using, e.g., interactions, memory, and coordination among control components) is processing a very special type of knowledge reflecting the agent perception of the hierarchy of needs (objectives, plans, etc.) and the current agent or the environment constraints. This kind of knowledge processing mainly deals with complex vague concepts (such as risk or safety) from the point of view of the selfish agent needs. Usually, this kind of knowledge processing is not necessarily logical reasoning in terms of proving statements (i.e., labeling statements by truth values such as TRUE or FALSE). This knowledge processing is rather analogous to the judgment process in a court aiming at recognition of evidence which could be used as an argument for or against. Arguments for or against are used in order to make the final decision which one of the solutions is the best for the agent in the current situation (i.e., arguments are labeling statements by judgment values expressing the action priorities). The evaluation of currents needs by agent Ag is realized from the point of view of hierarchy of agent Ag life values/needs). Wisdom type of knowledge processing by the agent Ag is characterized by the ability to improve quality of the judgment process based on the agent Ag experiences. In order to emphasize the importance of this ability, we use the concept of adaptive judgment in the wisdom equation instead of just judgment. An agent who is able to perform adaptive judgment in the above sense, we simply call as a judge.

The adaptivity aspects are also crucial from the point of view of interactions (Goldin et al., 2006; Nguyen & Skowron, 2008). The need for adaptation follows, e.g., from the fact that complex vague concepts on the basis of which the judgment is performed by the agent Ag are approximated by classification algorithms (classifiers) which should drift in time following changes in data and represented knowledge.

Complete Chapter List

Search this Book: