Using Proximity Graph Cut for Fast and Robust Instance-Based Classification in Large Datasets

Algorithm 1 Greedy graph-based in NSW for multiclass classification.

 Input: - dataset index; - sample to be classified

 Result: label

  random vertex from ;

 ;

 repeat

  ;

  ;

  // closest to x neighbours of d

  ;

  ;

  ;

  // until we can‘t get closer to x

 until;

 return