A Novel Impurity Measuring Technique for Decision Tree Learning in Machine Learning
D. Mabuni
1D. Mabuni*, Department of Computer Science, Dravidian University, Kuppam, India.
Manuscript received on May 16, 2020. | Revised Manuscript received on May 19, 2020. | Manuscript published on June 10, 2020. | PP: 506-512 | Volume-9 Issue-8, June 2020. | Retrieval Number: H6540069820/2020©BEIESP | DOI: 10.35940/ijitee.H6540.069820
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Decision tree classification is one of the most powerful data classification techniques in machine learning, data mining, big data analytics and split functionality is a crucial and inherently associated integral part of the decision tree learning. Many split similarity measures are proposed to determine the best split attribute and then partitioning the node data in decision tree learning accordingly. A new impurity measuring based split technique called (IMDT) for decision tree learning is proposed in this paper and it is used in obtaining experimental results. Many UCI machine learning dataset are employed in experimentation. The algorithm C4.5 is the most using data classification algorithm. The results obtained with the proposed approach are outperformed than the many existing decision tree classification algorithms in particular C4.5 decision tree algorithm.
Keywords: Big Data Analytics, Impurity Measure, Machine Learning, Split Functionality and Split Attribute.
Scope of the Article: Machine Learning