A Comprehensive and Comparative Study on Hierarchical Multi Label Classification
Purvi Prajapati1, Amit Thakkar2, Amit Ganatra3
1Purvi Prajapati, Department of Information Technology, Charotar University of Science and Technology, Changa , Anand, Gujarat, India.
2Amit Thakkar, Department of Information Technology, Charotar University of Science and Technology, Changa, Anand, Gujarat, India.
3Amit Ganatra, U and P U Patel Department of Computer Engineering, Charotar University of Science and Technology, Changa, Anand, Gujarat, India.
Manuscript received on January 17, 2012. | Revised Manuscript received on February 05, 2012. | Manuscript published on February 29, 2012. | PP: 110-116 | Volume-1 Issue-3, February 2012. | Retrieval Number: C0204021312/2011©BEIESP
Open Access | Ethics and Policies | Cite
© 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: Multi label classification is variation of single label classification where each instance is associated with more than one class labels. Multi label classification is used in many applications like text classification, gene functionality, image processing etc. Hierarchical multi-label classification problems combine the characteristics of both hierarchical and multi-label classification problems. This paper introduced k binary classifier and one classifier approaches of hierarchical multi label classification. These approaches are explained with two algorithms to solve hierarchical multi label classification problems. One is the C4.5H algorithm (extension of multi label decision tree) and second is Predictive Clustering Tree (PCT) algorithm. From theoretical and experimental study on yeast data set shows that PCT algorithm is the best option for hierarchical multi label classification. PCT algorithm is implemented on Clus. This paper introduced three approaches of Clus: Single Classification (SC), Hierarchical Single Label Classification (HSC) and Hierarchical Multi label Classification (HMC). From theoretical and experimental study, HMC performs better compare to remaining two approaches.
Keywords: Classification, Decision Tree, Hierarchical Classification, Multi Label Classification, Predictive clustering tree.