International Scholarly Research Notices

Research Article

Online Boosting Algorithm Based on Two-Phase SVM Training

Algorithm 1

The Pegasos algorithm with weighted samples.
function PEGASOS  ( β†’ 𝛼 , πœ† , β†’ π‘₯ , 𝑦 , 𝑑 , 𝑀 )
  𝑙 = m a x ( 0 , 1 βˆ’ ⟨ β†’ 𝛼 , β†’ π‘₯ ⟩ 𝑦 )
  𝜎 = 𝐼 ( 𝑙 > 0 )
  πœ‚ = 1 / πœ† 𝑑
  β†’ 𝛼 = ( 1 βˆ’ πœ‚ πœ† ) β†’ 𝛼 + 𝑀 𝜎 πœ‚ 𝑦 β†’ π‘₯
  β†’ 𝛼 = m i n ( 1 , 1 / ( β€– 𝑀 β€– 2 √ πœ† ) ) β†’ 𝛼
end function