Action Rules Mining in Hoarseness Disease

Action Rules Mining in Hoarseness Disease

Agnieszka Dardzinska (Bialystok University of Technology, Poland)
Copyright: © 2017 |Pages: 6
DOI: 10.4018/978-1-5225-0536-5.ch001


Action rule is an implication rule that shows the expected change in a decision value of an object as a result of changes made to some of its conditional values. An example of an action rule is ‘patients are expected to control their health regularly if they receive an information about free medical tests once a year'. In this case, the decision value is the health status, and the condition value is whether the information is sent to the patient. Because of some complex medical problems this paper discusses a strategy which generates action rules to using new knowledge base consisting of classification rules. As one of the testing domains for our research, we take new system for gathering and processing clinical data on patients with throat disorders, and mining action rules will suggest in simply way how to construct the decision support module for easier given diagnosis for patients.
Chapter Preview


Finding useful rules is an important task of a knowledge discovery process. Most researchers mainly focus on techniques for generating classification rules. A need for new methods with the ability to assist users in analyzing a large number of rules for a useful knowledge (Dardzinska, 2013) is still seeking. All patients, called objects, together with many symptoms and laboratory results, called attributes, form so called information When additional attributes, describing e.g. the situation of a patient are given, this system is called a decision information system. In such case, each object can be classified into one of the several given groups. An action rule is a rule extracted from a decision system, which gives suggestions helpful in reclassification process of objects in given information system from one state to another with respect to a distinguished attribute called a decision attribute (Ras, 2006). We assume that attributes are partitioned into stable (they cannot be changed, e.g. sex, name, height) and flexible (they can change, e.g. blood pressure, level of hoarseness). In paper (Ras,2008,2009) a new subclass of attributes called semistable attributes was introduced. They are typically a function of time, and undergo deterministic changes. It was shown in (Ras, 2008; Dardzinska, 2013) that some semistable attributes can be treated the same way as flexible attributes.

Complete Chapter List

Search this Book: