The Optimized Function Selection Using Wolf Algorithm for Classification

DOI

説明

Several classification techniques have been widely explored during the past decade. One of the novel approaches in recent years is Nature Based Algorithm. This approach is appropriate to imbalanced dataset. The focus of Nature Based Algorithm mostly is related to selection the optimized functions for self-learning. This is used to solve the NPhard problems. However, Some Nature Based Algorithms are suitable for general situation; some may be suitable for customized situation. This research proposes the Featured-Wolf (F-Wolf) algorithm to optimize the function selection problem in classification. The proposed algorithm applies the movement of a wolf and characteristics of wolves' leaders which can be more than one leader in a pack. Therefore, the pack can have more than one dominant leader which can help to select the most optimized functions to selection the most relevant features in the dataset. The experiment shows the comparison among other popular Nature Based Algorithms such as Ant Colony Optimization and other classification techniques. The results show that F-Wolf performs better results in terms of accuracy rate.

収録刊行物

キーワード

詳細情報 詳細情報について

  • CRID
    1390282763119456000
  • DOI
    10.5954/icarob.2017.gs10-1
  • ISSN
    21887829
  • 本文言語コード
    en
  • データソース種別
    • JaLC
    • Crossref
  • 抄録ライセンスフラグ
    使用不可

問題の指摘

ページトップへ