• DocumentCode
    3539937
  • Title

    Prediction of residential households´ water leakage using consensus method

  • Author

    Ismail, Norul Fadhilah ; Rasmani, Khairul A. ; Shahari, N. ; Rashid, Nur Rasyida Mohd ; Hanif, Herma Mohd ; Noh, Nor Aishah Mohd

  • Author_Institution
    Dept. of Math., Univ. Teknol. MARA Negeri Sembilan Campus, Kuala Pilah, Malaysia
  • fYear
    2012
  • fDate
    14-15 Aug. 2012
  • Firstpage
    50
  • Lastpage
    53
  • Abstract
    Consensus method is a means of communication between experts who assist the formation of a group judgment. This technique has great potential to be adopted to provide prediction based on outcomes obtained from several classification algorithms. In this study, data on water consumption was used to induce the classification model that will be used to predict the possibility of the occurrence of water leakage at residential premises. The classification algorithms used in the analysis using consensus method were selected based on their generalization ability on simulation datasets. The prediction outcomes were obtained based on analysis on individual algorithm classification outcomes on validation dataset. The finding showed a very promising result where consensus method produces consistent prediction outcomes.
  • Keywords
    classification; water supply; classification algorithm; classification model; consensus method; generalization ability; group judgment; residential household water leakage; simulation dataset; water consumption; Accuracy; Analytical models; Classification algorithms; Data models; Prediction algorithms; Predictive models; Training; classification algorithms; consensus method; residential households´ water leakage;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Uncertainty Reasoning and Knowledge Engineering (URKE), 2012 2nd International Conference on
  • Conference_Location
    Jalarta
  • Print_ISBN
    978-1-4673-1459-6
  • Type

    conf

  • DOI
    10.1109/URKE.2012.6319581
  • Filename
    6319581