Probabilistic Latent Semantic Analysis for Multichannel Biomedical Signal Clustering

Jin Wang, Mary She

    Research output: Contribution to journalArticlepeer-review

    6 Citations (Scopus)


    This letter extends probabilistic latent semantic analysis (pLSA) for multichannel biomedical signal clustering. The proposed multichannel pLSA (M-pLSA) models a multichannel signal as a generative process of local segments. It directly represents a biomedical signal as a mixture of latent topics based on the assumption that local segments extracted from each channel are conditionally independent given the topics. The categories of biomedical signals are automatically discovered in an unsupervised way. Experimental results demonstrate that the proposed M-pLSA model outperforms previous state-of-the-art methods and is robust to noise contamination.

    Original languageEnglish
    Article number7728039
    Pages (from-to)1821-1824
    Number of pages4
    JournalIEEE Signal Processing Letters
    Issue number12
    Publication statusPublished - 1 Dec 2016


    Dive into the research topics of 'Probabilistic Latent Semantic Analysis for Multichannel Biomedical Signal Clustering'. Together they form a unique fingerprint.

    Cite this