Multi-Complexity Measures for Early Detection and Monitoring of Neurological Abnormalities from Gait Time Series

摘要

Recently, we have proposed to use complementary complexity measures discovered by boosting-like ensemble learning for the enhancement of quantitative indicators dealing with necessarily short physiological time series. We have confirmed robustness of such multicomplexity measures for heart rate variability analysis with the emphasis on detection of emerging and intermittent cardiac abnormalities. Here we demonstrate that such ensemble-based approach could be also effective in discovering universal meta-indicators for early detection and convenient monitoring of neurological abnormalities using gait time series.