Time-Series Classification with Multivariate Statistical Dependence Features
This work addresses non-stationary time-series analysis for domains like speech recognition, offering a robust alternative to existing methods, though it appears incremental as it builds on FMCA.
The paper tackled time-series classification by introducing a framework that uses cross density ratio (CDR) to estimate statistical dependence, replacing correlation-based methods, and achieved higher accuracy on the TI-46 digit speech corpus with fewer than 10 layers and under 5 MB storage.
In this paper, we propose a novel framework for non-stationary time-series analysis that replaces conventional correlation-based statistics with direct estimation of statistical dependence in the normalized joint density of input and target signals, the cross density ratio (CDR). Unlike windowed correlation estimates, this measure is independent of sample order and robust to regime changes. The method builds on the functional maximal correlation algorithm (FMCA), which constructs a projection space by decomposing the eigenspectrum of the CDR. Multiscale features from this eigenspace are classified using a lightweight single-hidden-layer perceptron. On the TI-46 digit speech corpus, our approach outperforms hidden Markov models (HMMs) and state-of-the-art spiking neural networks, achieving higher accuracy with fewer than 10 layers and a storage footprint under 5 MB.