core algorithms and data structure

• quick sort
  sort samples along feature values
• shuffle
  random to draw a feature from the candidated features to sort
• dfs and bfs
  build a tree using deep first algorithm
• linked list + array = tree data structure
  using ‘array[node]’ data structure to represents a tree.
• row and column index array
  row[start:end] represents samples to select, and after sort only swap row[index_bigger] and row[index_litter] that is we only swap row[index] value and do not change index value, then we can pay attention only to index in range [start:end].