An Adaptive Ensemble Classifier for Mining Concept-Drifting Data Streams

UIU Institutional Repository

    • Login
    View Item 
    •   UIU DSpace Home
    • School of Science and Engineering (SoSE)
    • Department of Computer Science and Engineering (CSE)
    • Faculty Publications
    • View Item
    •   UIU DSpace Home
    • School of Science and Engineering (SoSE)
    • Department of Computer Science and Engineering (CSE)
    • Faculty Publications
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    An Adaptive Ensemble Classifier for Mining Concept-Drifting Data Streams

    Thumbnail
    View/Open
    An Adaptive Ensemble Classifier for Mining Concept-Drifting Data Streams.pdf (237.4Kb)
    Date
    2013-11-01
    Author
    Rahman, Chowdhury Mofizur
    Farid, Dewan Md
    Zhang, Li
    Hossain, Alamgir
    Strachan, Rebecca
    Sexton, Graham
    Dahal, Keshav
    Metadata
    Show full item record
    Abstract
    Traditional data mining techniques cannot be directly applied to the real-time data streaming environment. Existing mining classifiers therefore need to be updated frequently to adopt the changes in data streams. In this paper, we address this issue and propose an adaptive ensemble approach for classification and novel class detection in concept-drifting data streams. The proposed approach uses traditional mining classifiers and updates the ensemble model automatically so that it represents the most recent concepts in data streams. For novel class detection we consider the idea that data points belonging to the same class should be closer to each other and should be far apart from the data points belonging to other classes. If a data point is well separated from the existing data clusters, it is identified as a novel class instance. We tested the performance of this proposed stream classification model against existing mining algorithms using real benchmark datasets from UCI machine learning repository. The experimental results proved that our approach shows great flexibility and robustness in novel class detection in concept-drifting and outperforms traditional classification models in challenging real-life data stream applications.
    URI
    http://dspace.uiu.ac.bd/handle/52243/72
    Collections
    • Faculty Publications [11]

    Copyright 2003-2017 United International University
    Contact Us | Send Feedback
    Developed by UIU CITS
     

     

    Browse

    All of DSpaceCommunities & CollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

    My Account

    LoginRegister

    Copyright 2003-2017 United International University
    Contact Us | Send Feedback
    Developed by UIU CITS