[BACK]

Word composition metrics are based in L-tuple frequencies or counts. Distances between two sequences are defined as:

•  Smith-Waterman alignment algorithm (SW)
  Smith-Waterman raw score, using BLOSUM50 scoring matrix and a linear scoring scheme with gap penalty of 8
• W-metric (Wm)
  see manuscript
• Mahalanobis (ma)
  
• Standard Euclidean (se)
  
• Euclidean (eu)
  
• Kullback-Leibler discrepancy (ku)
  
• Cosine (co)
  
  , where 
                             
  

For further information and definitions see (Vinga & Almeida, 2003)

[BACK]