Research Article

Method for Solving LASSO Problem Based on Multidimensional Weight

Pseudocode 1

Input. Response dependent on predictor , error tolerance
Output. LASSO solution for (3)
() begin
() data preprocessing: Normalization ()
() initialization: ,
()
()
() ,
() calculating weight by     or    or  
() centralization ;
() when and do cycle;
() in LARS cycle
      
      
() end cycle;
() return .