A Feature Selection Method for Multi-Dimension Time-Series Data
This addresses feature redundancy in time-series applications like wearable sensors or video analysis, offering a more efficient selection method, though it appears incremental as it builds on existing mutual information and classifier-based approaches.
The paper tackles the problem of feature selection for multi-dimensional time-series data, such as from motion capture or activity recognition, by proposing a mutual information-based method (MSTS) that calculates a merit score from classifier outputs; it shows MSTS is significantly more computationally efficient while maintaining good accuracy compared to Wrapper-based methods, as evaluated on six datasets.
Time-series data in application areas such as motion capture and activity recognition is often multi-dimension. In these application areas data typically comes from wearable sensors or is extracted from video. There is a lot of redundancy in these data streams and good classification accuracy will often be achievable with a small number of features (dimensions). In this paper we present a method for feature subset selection on multidimensional time-series data based on mutual information. This method calculates a merit score (MSTS) based on correlation patterns of the outputs of classifiers trained on single features and the `best' subset is selected accordingly. MSTS was found to be significantly more efficient in terms of computational cost while also managing to maintain a good overall accuracy when compared to Wrapper-based feature selection, a feature selection strategy that is popular elsewhere in Machine Learning. We describe the motivations behind this feature selection strategy and evaluate its effectiveness on six time series datasets.