ELMVIS: a Nonlinear Visualization Technique using Random Permutations and Extreme Learning Machine (to appear)
AUTHORS:
Akusok Anton,
Miche Yoan
,
Lendasse Amaury
,
Nian Rui
JOURNAL:
IEEE Intelligent Systems
VOLUME:
NA
PAGES:
NA
@article{ Akusok2013a, author = "Akusok, Anton and Miche, Yoan and Lendasse, Amaury and Nian, Rui", responsibleauthor = "Anton, Akusok and Miche, Yoan", language = "eng", title = "ELMVIS: a Nonlinear Visualization Technique using Random Permutations and Extreme Learning Machine", url = "", country = "United States", abstract = "A nonlinear data visualization method based on extreme learning machine (ELM) is proposed in this paper. The novel idea is to pair original data points with pre-selected visualization ones, using random permutations and the adjusted ELM as an extremely fast nonlinear error metric. Despite the fact that the exact solution with permutation has a factorial complexity, experiments show a fast convergence rate for small to medium sample sets. The method is tested on three benchmark datasets, namely a spiral, artificial and real face images datasets. The proposed methodology succeeded in all of them, showing performance comparable or above the current state-of-the-art visualization methods.", issn = "NA", number = "NA", pages = "NA", volume = "NA", juforank = "2", flags = "AIRC", il = "yes", keywords = "ELM, random permutation, visualization", year = "to appear", date = "2013-08-05", unitcode = "T306-100", kay = "NA", impactfactor = "A1", journal = "IEEE Intelligent Systems" }