A benchmark of data stream classification for human activity recognition on connected objects
This work addresses the problem of efficient and accurate human activity recognition on resource-constrained connected devices, but it is incremental as it benchmarks existing methods without introducing new algorithms.
The paper evaluated five data stream classifiers for human activity recognition on connected devices, finding that Hoeffding Tree, Mondrian Forest, and Naive Bayes outperformed others on most datasets but still performed substantially worse than offline classifiers, with high memory consumption and low F1 scores as key challenges.
This paper evaluates data stream classifiers from the perspective of connected devices, focusing on the use case of HAR. We measure both classification performance and resource consumption (runtime, memory, and power) of five usual stream classification algorithms, implemented in a consistent library, and applied to two real human activity datasets and to three synthetic datasets. Regarding classification performance, results show an overall superiority of the HT, the MF, and the NB classifiers over the FNN and the Micro Cluster Nearest Neighbor (MCNN) classifiers on 4 datasets out of 6, including the real ones. In addition, the HT, and to some extent MCNN, are the only classifiers that can recover from a concept drift. Overall, the three leading classifiers still perform substantially lower than an offline classifier on the real datasets. Regarding resource consumption, the HT and the MF are the most memory intensive and have the longest runtime, however, no difference in power consumption is found between classifiers. We conclude that stream learning for HAR on connected objects is challenged by two factors which could lead to interesting future work: a high memory consumption and low F1 scores overall.