Multi-Kernel NLMS Algorithm with Coherence Criterion and Its Application to Online Prediction of Time Series Data
スポンサーリンク
概要
- 論文の詳細を見る
In this paper, we propose a nonlinear adaptive filtering algorithm using multiple kernels. The proposed algorithm is a generalization of the kernel normalized least mean square (KNLMS) algorithm (Richard et al., 2009). The algorithm utilizes a coherence criterion for dictionary selection. Numerical examples show the advantages of the proposed algorithm in online prediction of nonstationary time series data.
- 2011-05-05