Terahertz Frequency Domain Spectroscopy Identification System Based on Decision Trees

The application of pattern recognition methodology within chemistry, biology and other science domains, especially in security systems is becoming more and more important. Many classification algorithms are available in literature but decision trees are the most commonly exploited because of their ease of implementation and understanding in comparison to other classification algorithms. Decision trees are powerful and popular tools for classification and prediction. In contrast to neural networks, decision trees represent rules, which can readily be expressed so that humans can understand them or even directly use in a database. In this paper we present an algorithm of construction of decision trees and a classification rule extraction based on a logical relationship between attributes and a generalized decision function. Moreover, correctness and efficiency of the algorithm was experimentally validated in a terahertz system, where spectra of explosives were measured in reflection configuration.