Skip to content

A MATLAB implementation of Adaptive k-Nearest Neighbor Algorithms called Ada-kNN and Ada-kNN2 (alongside a global weighting scheme for handling class imbalance).

License

Notifications You must be signed in to change notification settings

mlrgisi/Ada-kNN

 
 

Repository files navigation

A MATLAB implementation of Ada-kNN, Ada-kNN2, Ada-kNN+GIHS and Ada-kNN2+GIHS.

Written by: Sankha Subhra Mullick.

Reference: S. S. Mullick, S. Datta and S. Das, "Adaptive Learning-Based k-Nearest Neighbor Classifiers With Resilience to Class Imbalance," in IEEE Transactions on Neural Networks and Learning Systems, doi: 10.1109/TNNLS.2018.2812279.

Contact: [email protected] (Sankha Subhra Mullick).

DESCRIPTION:

  • The package contains 7 functions.
  • adaKnn.m: Function implementing Ada-kNN algorithm.
  • adaKnn2.m: Function implementing Ada-kNN2 algorithm.
  • adaKnnGIHS.m: Function implementing Ada-kNN coupled with GIHS algorithm (for imbalanced classification).
  • adaKnn2GIHS.m: Function implementing Ada-kNN coupled with GIHS algorithm (for imbalanced classification).
  • kNNIMB.m: Function implementing weighted k-nearest neighbor algorithm.
  • learningModel.m: Function implementing the proposed heuristic learning technique (to be used by adaKnn2 and adaKnn2GIHS).
  • standardised.m: Supporting function.

DEPENDENCIES:

  • MATLAB 2014a and above.
  • Neural Network toolbox (for adaKnn and adaKnnGIHS).

About

A MATLAB implementation of Adaptive k-Nearest Neighbor Algorithms called Ada-kNN and Ada-kNN2 (alongside a global weighting scheme for handling class imbalance).

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • MATLAB 100.0%