Parallel confidence-weighted classification of large-scale, multimodal neural data on MapReduce
-
1
University of Minnesota, United States
Existing neuroimaging technologies, such as magnetic resonance imaging and magnetoencephalography, continue to produce large, complex, high-dimensional data for numerous psychological disorders and neurological diseases. To identify these disorders/diseases from neural data, researchers typically apply one classification algorithm or compare multiple classification algorithms. By linking multiple classifiers (e.g. k-Nearest Neighbors, Naïve Bayesian, Decision Tree, k-Means, Support Vector Machines, Expectation-Maximization, etc.), the collective performance of the various algorithms can be exploited to yield consistent overall classification performance. Here we introduce the linked-classifier, which processes weighted inputs of several classifiers to yield a final, confidence-based classification outcome. Now, the application of multiple classification algorithms on large-scale data creates both memory and time constraints on standard, stand-alone computers. We overcome this problem by implementing our analysis within a MapReduce framework (see Figure 1) running in a computer cluster. The MapReduce framework is a computing paradigm that simplifies data demanding parallel computation. Thus, by taking advantage of this framework and the speed of its implementation, the multi-linked classifier functions run fast allowing for efficient robust classifications.
Keywords:
MapReduce,
MRI imaging,
Magnetoencephalography,
Classification,
machine learning
Conference:
Neuroinformatics 2014, Leiden, Netherlands, 25 Aug - 27 Aug, 2014.
Presentation Type:
Poster, to be considered for oral presentation
Topic:
General neuroinformatics
Citation:
Mahan
MY and
Georgopoulos
A
(2014). Parallel confidence-weighted classification of large-scale, multimodal neural data on MapReduce.
Front. Neuroinform.
Conference Abstract:
Neuroinformatics 2014.
doi: 10.3389/conf.fninf.2014.18.00078
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
28 Apr 2014;
Published Online:
04 Jun 2014.
*
Correspondence:
Ms. Margaret Y Mahan, University of Minnesota, Minneapolis, United States, 685785@frontiersin.org