A Survey on Classification Methods Based on Decision Tree Algorithms in Data Mining

Abstract

Data mining resides in the junction of traditional statistics and computer science. As distinct from statistics, data mining is more about searching for hypotheses in data that happens to be available instead of verifying research hypotheses by collecting data from designed experiments. Data mining is also characterized as being oriented toward problems with a large number of variables and/or samples that makes scaling up algorithms important. This means developing algorithms with low computational complexity, using parallel computing, partitioning the data into subsets, or finding effective ways to use relational data bases. The process- and utility-centered thinking in data mining and knowledge discovery is manifested also in the reported, commercial systems. Decision Trees are considered to be one of the most popular approaches for representing classifiers. Researchers from various disciplines such as statistics, machine learning, pattern recognition, and data mining considered the issue of growing a decision tree from available data. The technology for building Knowledge based system by decision tree algorithms has been demonstrated successfully in several practical applications. This paper summarizes an approach to synthesizing decision trees that has been used in variety of systems, and it describes such system ID3, C4.5 and CART. Results from recent studies show ways in which the methodology can be modified to deal with information that is noisy and/or incomplete.