Enhanced Decision Tree Algorithm for Discovery of Exceptions

Decision trees are the most admired and extensively used classification algorithms in data mining. These are considered accurate, easy to use, and comprehensible classifiers. Like many other classification models, decision tree classifiers ignore exceptions as noise. Exceptions are precious pieces of knowledge that are required to modify our decisions in extraordinarily rare circumstances. Since exceptions pertain to a tiny number of instances, it is a very challenging task to capture exceptions since a learning algorithm’s focus is on discovering knowledge with high generalization power. This paper proposes an enhanced decision tree learning algorithm that accommodates exceptions. The working of the suggested algorithm is demonstrated on a toy example dataset. It is further applied to three datasets obtained from machine learning repository. The results reveal that the Enhanced Decision Tree Algorithm (EDTA) discovers several exceptions across the experimental datasets.

[1]  Kamal Kant Bharadwaj.  A parellel genetic algorithm approach for automated discovery of censored production rules , 2007, Artificial Intelligence and Applications.

[2]  Wei-Yin Loh,et al.  A Comparison of Prediction Accuracy, Complexity, and Training Time of Thirty-Three Old and New Classification Algorithms , 2000, Machine Learning.

[3]  Renu Bala.  DISCOVERING FUZZY CEN SORED CLASSIFICATION RULES (FCCRS ):A GENETIC ALGORITHM APPROACH , 2012 .

[4]  Onur Dikmen,et al.  Parallel univariate decision trees , 2007, Pattern Recognit. Lett..

[5]  Kamal K. Bharadwaj,et al.  Evolutionary Approach for Automated Discovery of Censored Production Rules , 2007 .

[6]  Jorma Rissanen,et al.  SLIQ: A Fast Scalable Classifier for Data Mining , 1996, EDBT.

[7]  Jan M. Zytkow,et al.  Unified Algorithm for Undirected Discovery of Execption Rules , 2000, PKDD.

[8]  Thomas G. Dietterich.  Multiple Classifier Systems , 2000, Lecture Notes in Computer Science.

[9]  Harry Zhang,et al.  A Fast Decision Tree Learning Algorithm , 2006, AAAI.

[10]  Ashwin Srinivasan,et al.  Ripple down rules: Turning knowledge acquisition into knowledge maintenance , 1992, Artif. Intell. Medicine.

[11]  S. Appavu alias Balamurugan,et al.  Effective solution for unhandled exception in decision tree induction algorithms , 2009, Expert Syst. Appl..

[12]  Rakesh Agrawal,et al.  SPRINT: A Scalable Parallel Classifier for Data Mining , 1996, VLDB.

[13]  Anthony J. Bagnall,et al.  Improving Decision Tree Performance Through Induction- and Cluster-Based Stratified Sampling , 2004, IDEAL.

[14]  Howard J. Hamilton,et al.  Interestingness measures for data mining: A survey , 2006, CSUR.

[15]  Sotiris B. Kotsiantis,et al.  Decision trees: a recent overview , 2011, Artificial Intelligence Review.

[16]  Deborah R. Carvalho,et al.  A hybrid decision tree/genetic algorithm method for data mining , 2004, Inf. Sci..

[17]  Renu Bala,et al.  A Genetic Algorithm Approach for Discovering Tuned Fuzzy Classification Rules with Intra- and Inter-Class Exceptions , 2016, J. Intell. Syst..

[18]  Dharminder Kumar,et al.  An evolutionary approach to discover intra- and inter-class exceptions in databases , 2013, Int. J. Intell. Syst. Technol. Appl..

[19]  S. Appavu alias Balamurugan,et al.  Improving decision tree performance by exception handling , 2010, Int. J. Autom. Comput..

[20]  JOHANNES GEHRKE,et al.  RainForest—A Framework for Fast Decision Tree Construction of Large Datasets , 1998, Data Mining and Knowledge Discovery.