Performance evaluation of decision tree versus artificial neural network based classifiers in diversity of datasets

Large databases of digital information are ubiquitous. Data from the neighborhood store´s checkout register, your bank´s credit card authorization device, records in your doctor´s office, patterns in your telephone calls and many more applications generate streams of digital records archived in huge databases, sometimes in so-called data warehouses A new generation of computational techniques and tools is required to support the extraction of useful knowledge from the rapidly growing volumes of data. These techniques and tools are the subject of the emerging field of knowledge discovery in databases (KDD) and data mining. Data mining plays an important role to discover important information to help in decision making of a decision support system. It has been the active area of research in the last decade. The classification is one of the important tasks of data mining. Different kind of classifiers have been suggested and tested to predict the future events based on unseen data. This paper compares the performance evaluation of decision tree and artificial neural network based classifiers in diversity of datasets. Three decision trees (CHAID, QUEST and C5.0), and one ANN based back propagation classifier have been compared in terms of predictive accuracy, training time and comprehensibility. Out of the decision trees, QUEST generates trees with lesser levels and depth showing more comprehensibility. C5.0 and back propagation classifier show predictive accuracy of the same order. Back propagation based classifier shows zero comprehensibility. This research work shows that decision tree based classifiers are better for organizational decision support systems as compared to ANN based classifiers.