cv52                   package:gcl                   R Documentation

_5_x_2 _f_o_l_d _c_r_o_s_s_v_a_l_i_d_a_t_i_o_n

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

     Function to perform 5x2-fold crossvalidation comparison of two
     predictive models on the same data.

_U_s_a_g_e:

     cv52(df, mb1, mb2, eval, seed = NULL, cv.verbose = F, ...)

_A_r_g_u_m_e_n_t_s:

      df: The data frame, last column is outcome

     mb1: Model builder 1, must return a function that takes 'df' as
          input and returns a matrix where entry (i,j) is the
          classifiers belief of case i's memebership in class j

     mb2: Model builder 2, must return a function that takes 'df' as
          input and returns a matrix where entry (i,j) is the
          classifiers belief of case i's memebership in class j

    eval: Evaluation function, takes as input classifier generated by
          mb[12] above and data frame 'df'

    seed: Random seed.

cv.verbose: If TRUE, spew info.

     ...: Passed on.

_D_e_t_a_i_l_s:

     This function implements the 5x2 CV F statistic proposed by
     Alpaydin as an improvement of Dietterich's statistic.

_V_a_l_u_e:

     Returns a list. 

       f: the F statistic value

       p: p value

       m: matrix of split/fold results

   seeds: vector of random seeds used

_A_u_t_h_o_r(_s):

     Staal A. Vinterbo staal@dsg.harvard.edu

_R_e_f_e_r_e_n_c_e_s:

     Alpaydin: "Combined 5x2 cv F test for comparing supervised
     classification learning algorithms" Neural Computation 1999

_S_e_e _A_l_s_o:

     'cv','acc.eval','ci.eval'

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

     ## Not run: 
     cv52(df, gcl, tcl, acc.eval)
     ## End(Not run)

