Multi-label methods for prediction with sequential data. (March 2017)

• Record Type: Journal Article
• Authors: Read, Jesse; Martino, Luca; Hollmén, Jaakko
• Is Part Of: Pattern recognition. Volume 63(2017:Mar.)
• Page Start: 45
• Page End: 55
• Publication Date: 2017-03
• Subjects: Multi-label classification -- Problem transformation -- Sequential data -- Sequence prediction -- Markov models; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2016.09.015
• Languages: English
• ISSNs: 0031-3203

Abstract

Abstract: The number of methods available for classification of multi-label data has increased rapidly over recent years, yet relatively few links have been made with the related task of classification of sequential data. If labels indices are considered as time indices, the problems can often be seen as equivalent. In this paper we detect and elaborate on connections between multi-label methods and Markovian models, and study the suitability of multi-label methods for prediction in sequential data. From this study we draw upon the most suitable techniques from the area and develop two novel competitive approaches which can be applied to either kind of data. We carry out an empirical evaluation investigating performance on real-world sequential-prediction tasks: electricity demand, and route prediction. As well as showing that several popular multi-label algorithms are in fact easily applicable to sequencing tasks, our novel approaches, which benefit from a unified view of these areas, prove very competitive against established methods. Abstract : Highlights: Drawing connections between learning for multi-label and sequential data. A unified view between multi-label and sequential classifiers. A novel Markov model-inspired method for multi-label (and sequence) classification. A novel multi-label-inspired method for sequence (and multi-label) classification. An empirical comparison with related methods, on real-world datasets, demonstrating the competitiveness of proposed methods.