CHAPTER 2
FUNDAMENTALS OF PATTERN ANALYSIS: A BRIEF OVERVIEW
2.1 INTRODUCTION
With the rapid proliferation of the use of computers, a vast amount of data are generated in every area of engineering and scientific disciplines (biology, psychology, medicine, marketing, finance, etc.). This vast amount of data with potentially useful information needs to be analyzed automatically for extraction of hidden knowledge. Pattern analysis is the process of automatically detecting patterns characterizing the inherent information in data. A pattern is defined in [1] as opposite of chaos; It is an entity, vaguely defined, that could be given a name. The objective of pattern analysis is to identify the patterns into some known categories–classes or to group the patterns in different categories that are then assigned some tags–class names. The former is known as supervised pattern classification and the later is known as unsupervised pattern classification or clustering.
The area of pattern analysis or pattern classification is not a new one, the research began during the 1950s and varieties of techniques [2–4] have been developed over the years. The survey papers of Unger [5], Wee [6], and Nagy [7] represent an account of the earlier works done in this area. Classical data analysis techniques for pattern discovery are based mainly on statistics and mathematics [8–10]. An excellent overview of statistical pattern recognition techniques is available in [11]. Though statistical ...
