New Measures for Constructing Decision Tree Based on Rough Sets and Applications

Nguyen Duc Thuan, Pham Quang Tung

Abstract


In a decision tree algorithm, an important factor is measure for selecting the best split the records. There are many measures that can
be used to determine the splitting attributes based on Entropy, Gini index, Information Gain and so on. In this paper, two metrics from the
consistency degree of two knowledges is used as the criteria for selecting the attribute that will best separate the samples into individual classes.
These new measures also to define on covering rough sets for constructing decision tree on incomplete information systems.


Keywords: Decision tree, Rough set, Consistency degree, Splitting Attribute


Full Text:

PDF


DOI: https://doi.org/10.26483/ijarcs.v6i5.2504

Refbacks

  • There are currently no refbacks.




Copyright (c) 2016 International Journal of Advanced Research in Computer Science