Reducing multiclass to binary: a unifying approach for margin classifiers
The Journal of Machine Learning Research
Gesture recognition with a Wii controller
Proceedings of the 2nd international conference on Tangible and embedded interaction
Pattern Recognition, Fourth Edition
Pattern Recognition, Fourth Edition
MED '09 Proceedings of the 2009 17th Mediterranean Conference on Control and Automation
Activity classification using realistic data from wearable sensors
IEEE Transactions on Information Technology in Biomedicine
Hi-index | 0.00 |
Physical activity recognition has become an increasing research area specially on health-related fields. The amount of different postures, movements and exercises in addition to the difficulty of the individuals particular execution style determine that extremely robust efficient knowledge inference systems are extremely necessary, being classification process one of the most crucial steps. Considering the power of binary classification in contrast to direct multiclass approaches, and the capabilities offered by multi-sense environments, we define a novel classification schema based on hierarchical structures composed by weighted decision makers. Remarkable accuracy results are obtained for a particular activity recognition problem in contrast to a traditional multiclass majority voting algorithm.