Comparison of approaches to multilabel and multitarget classification

Saša Saftić (2013) Comparison of approaches to multilabel and multitarget classification. EngD thesis.

Preview

Abstract

In this research we compare dierent algorithms for single-label, multi-label and multi-class classication. The testing was performed using data representing a problem from educational domain, and are appropriate for multiclass classication. We used dierent methods for transforming multi-label classication to single-label classication. We compared algorithms with evaluation measures and ROC curves. Results shows that several methods can be used for this task. Algorithms that are specialized for multi-class classication have better predictions for multi-class problems, although other algorithms also show good performance.

Item Type:

Thesis (EngD thesis)

Keywords:

single-label classification, multi-label classification, multi-class classification, evaluation measures, ROC curve

Number of Pages:

Language of Content:

Slovenian

Mentor / Comentors:

Name and Surname	ID	Function
doc. dr. Zoran Bosnić	3826	Mentor

Link to COBISS:

http://www.cobiss.si/scripts/cobiss?command=search&base=50070&select=(ID=10244180)

Institution:

University of Ljubljana

Department:

Faculty of Computer and Information Science

Item ID:

2149

Date Deposited:

13 Sep 2013 15:47

Last Modified:

11 Nov 2013 14:58

URI:

http://eprints.fri.uni-lj.si/id/eprint/2149

Actions (login required)

View Item