varImpStruct-class       package:MLInterfaces       R Documentation

_C_l_a_s_s "_v_a_r_I_m_p_S_t_r_u_c_t" - _c_o_l_l_e_c_t _d_a_t_a _o_n _v_a_r_i_a_b_l_e _i_m_p_o_r_t_a_n_c_e
_f_r_o_m _v_a_r_i_o_u_s _m_a_c_h_i_n_e _l_e_a_r_n_i_n_g _m_e_t_h_o_d_s

_D_e_s_c_r_i_p_t_i_o_n:

     collects data on variable importance

_O_b_j_e_c_t_s _f_r_o_m _t_h_e _C_l_a_s_s:

     Objects can be created by calls of the form 'new("varImpStruct",
     ...)'. These are matrices of importance measures with separate 
     slots identifying algorithm generating the measures and variable
     names.

_S_l_o_t_s:

     '._D_a_t_a': Object of class '"matrix"' actual importance measures 

     '_m_e_t_h_o_d': Object of class '"character"' tag 

     '_v_a_r_n_a_m_e_s': Object of class '"character"' conformant vector of
          names of variables 

_E_x_t_e_n_d_s:

     Class '"matrix"', from data part. Class '"structure"', by class
     '"matrix"'. Class '"array"', by class '"matrix"'. Class
     '"vector"', by class "matrix", with explicit coerce. Class
     '"vector"', by class "matrix", with explicit coerce.

_M_e_t_h_o_d_s:

     _p_l_o_t 'signature(x = "varImpStruct")': make a bar plot, you can
          supply an argument resolveenv which will look up variable
          names in a translation environment 

     _s_h_o_w 'signature(object = "varImpStruct")': simple abbreviated
          display 

     _g_e_t_V_a_r_I_m_p 'signature(object = "classifOutput")': extractor of
          variable importance structure

_E_x_a_m_p_l_e_s:

     library(golubEsets)
     data(Golub_Merge)
     library(hu6800)
     smallG <- Golub_Merge[1:60,]
     set.seed(1234)
     opar=par(no.readonly=TRUE)
     par(las=2, mar=c(10,11,5,5))
     rf <- randomForestB( smallG, "ALL.AML", 1:40, importance=TRUE )
     plot( getVarImp( rf ), n=10, resolveenv=hu6800SYMBOL )
     gb <- gbmB( smallG, "ALL.AML", 1:40 , n.minobsinnode=3 , n.trees=6000)
     plot( getVarImp( gb ), n=10, resolveenv=hu6800SYMBOL )
     set.seed(1234)
     rf2 <- MLearn(ALL.AML~., smallG, randomForestI, 1:40, importance=TRUE,
      sampsize=table(smallG$ALL.AML[1:40]), mtry=sqrt(ncol(exprs(smallG))))
     plot( getVarImp( rf2 ), n=10, resolveenv=hu6800SYMBOL ) # translation?
     par(opar)

