NFEA

  张量分解提取生物学特征，NFEA: Tensor Toolbox for Feature Extraction and Applications( Data in modern applications such as BCI based on EEG signals often contain multi-modes due to mechanism of data recording, e.g. signals recorded by multiple-sensors (electrodes), in multiple trials, epochs, for multiple subjects and with different tasks, conditions. Moreover, during processing and analysis, dimensionality of the data could be augmented due to expression of the data into sparse domain (time-frequency representation) by different transforms such as STFT, wavelets. That means data itself is naturally a tensor, and has multilinear structures. Standard approaches which analyze such data by considering them as vectors or matrices might be not suitable due to risk of losing the covariance information among various modes. To discover hidden multilinear structures, features within the data, the analysis tools should reflect the multi-dimensional structure of the data)