Multi-label Methods for Prediction with Sequential Data

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. Keywords: multi-label classification; problem transformation; sequential data; sequence prediction; Markov models 1. Introduction Multi-label classification is the supervised learning problem where an instance is associated with multiple class variables (i.e., labels), rather than with a single class, as in traditional classification problems. See [1] for a review. Corresponding author, jesse.read@polytechnique.edu Preprint submitted to Pattern Recognition September 29, 2016 labels were modelled independently - at the expense of an increased computational cost. The case of binary labels is most common, where a positive class value denotes the relevance of the label (and the negative or null class denotes irrelevance). Typical examples of binary multi-label classification involve categorizing text documents and images, which can be assigned any subset of a particular label set. For example, an image can be associated with both labels beach and sunset. The multi-label classification paradigm has been successfully considered also in many other domains, such as text, video, audio, and bioinformatics - see [1] and references therein for further examples.