Title:

On Computational Complexity of Multiclass Classification Approach ECOC

Author:

Danoyan Hayk

Type:

Conference

Uncontrolled Keywords:

Computational efficiency ; ECOC ; Ensemble Methods ; Minimum Test Collection ; Multiclass Classification

Abstract:

The multiclass classification approach known as ECOC (error correcting output codes) is considered. The method solves the multiclass classification problem by combining binary classification algorithms according to some binary matrix. The framework is considered where the columns of the mentioned matrix are selected from the given set. It is proved that the algorithmic problem of column selection optimizing the training error is NP complete

Language:

English

URL:


Affiliation:

Institute for Informatics and Automation Problems

Country:

Armenia

Year:

2017

Time period:

September25-29

Conference title:

11th International Conference on Computer Science and Information Technologies CSIT 2017

Place:

Yerevan

Participation type:

oral