The arboretum procedure
Yüklə 3,07 Mb. Pdf görüntüsü
|
60 X2_H30 1.075612 0.0004997 X2 -> H30 61 BIAS_H1 1.652625 0.0002540 BIAS -> H1 62 BIAS_H2 0.729833 0.0007927 BIAS -> H2 63 BIAS_H3 -0.029132 -0.0002571 BIAS -> H3 64 BIAS_H4 -0.554034 0.0000857 BIAS -> H4 65 BIAS_H5 1.242018 0.0000550 BIAS -> H5 66 BIAS_H6 -0.429913 -0.00191 BIAS -> H6 67 BIAS_H7 0.171418 -0.0003334 BIAS -> H7 68 BIAS_H8 0.668461 0.0001098 BIAS -> H8 69 BIAS_H9 -0.985128 -0.00175 BIAS -> H9 70 BIAS_H10 0.362497 0.0000239 BIAS -> H10 71 BIAS_H11 -0.728020 -0.0002738 BIAS -> H11 72 BIAS_H12 0.742591 -0.0004105 BIAS -> H12 73 BIAS_H13 -1.142064 -0.0004769 BIAS -> H13 74 BIAS_H14 0.456503 0.0001368 BIAS -> H14 75 BIAS_H15 1.429261 0.00137 BIAS -> H15 76 BIAS_H16 -1.405924 0.00109 BIAS -> H16 77 BIAS_H17 -0.861128 -0.0000339 BIAS -> H17 78 BIAS_H18 -0.888192 -0.00480 BIAS -> H18 79 BIAS_H19 -1.509340 -0.00139 BIAS -> H19 80 BIAS_H20 0.431651 6.87045E-7 BIAS -> H20 81 BIAS_H21 -0.018080 -0.0000773 BIAS -> H21 82 BIAS_H22 -0.663759 7.53178E-7 BIAS -> H22 83 BIAS_H23 0.178667 0.00162 BIAS -> H23 84 BIAS_H24 0.515460 -0.0006172 BIAS -> H24 85 BIAS_H25 -0.721835 -0.0000149 BIAS -> H25 ---------------------------------------------------------- Parameter Estimate Gradient Label ---------------------------------------------------------- 86 BIAS_H26 -0.241293 0.0000243 BIAS -> H26 87 BIAS_H27 0.670665 0.0000766 BIAS -> H27 88 BIAS_H28 0.255718 0.0005082 BIAS -> H28 89 BIAS_H29 0.195161 0.0000364 BIAS -> H29 90 BIAS_H30 1.509699 -0.00366 BIAS -> H30 91 H1_HIPL -0.136499 -0.00160 H1 -> HIPL 92 H2_HIPL -0.091000 -0.00300 H2 -> HIPL 93 H3_HIPL -0.144627 -0.00174 H3 -> HIPL 94 H4_HIPL 0.175297 -0.00214 H4 -> HIPL 95 H5_HIPL -0.078282 -0.00190 H5 -> HIPL 96 H6_HIPL 0.803260 -0.00174 H6 -> HIPL 97 H7_HIPL -0.132805 0.0004925 H7 -> HIPL 98 H8_HIPL -0.057435 -0.0008912 H8 -> HIPL 99 H9_HIPL -0.333394 0.0005940 H9 -> HIPL 100 H10_HIPL -0.092374 -0.0004988 H10 -> HIPL 101 H11_HIPL -0.152927 0.00433 H11 -> HIPL 102 H12_HIPL 0.117821 -0.00294 H12 -> HIPL 103 H13_HIPL 0.206879 -0.00262 H13 -> HIPL 104 H14_HIPL -0.302093 0.00236 H14 -> HIPL 105 H15_HIPL -0.221776 0.00168 H15 -> HIPL 106 H16_HIPL -0.366827 -0.00201 H16 -> HIPL 107 H17_HIPL -0.062601 0.00624 H17 -> HIPL 108 H18_HIPL 0.693210 0.00477 H18 -> HIPL 109 H19_HIPL -0.438978 -0.00270 H19 -> HIPL 110 H20_HIPL 0.004246 -0.00201 H20 -> HIPL 111 H21_HIPL -0.195547 0.00236 H21 -> HIPL 112 H22_HIPL 0.004032 0.00203 H22 -> HIPL 113 H23_HIPL 0.307420 -0.00475 H23 -> HIPL 114 H24_HIPL 0.391365 0.00300 H24 -> HIPL 115 H25_HIPL 0.051031 0.00192 H25 -> HIPL 116 H26_HIPL -0.041873 0.00200 H26 -> HIPL 117 H27_HIPL -0.061313 0.00141 H27 -> HIPL 118 H28_HIPL -0.285679 -0.00139 H28 -> HIPL 119 H29_HIPL -0.118849 -0.00205 H29 -> HIPL 120 H30_HIPL 0.374561 -0.00512 H30 -> HIPL 121 BIAS_HIP -0.623399 -0.00230 BIAS -> HIPL Hill & Plateau Data MLP with 30 Hidden Units Dual Quasi-Newton Optimization *** Termination Criteria *** Minimum Iterations . . . . . . . . . . . . . . . -1 Maximum Iterations . . . . . . . . . . . . . . . 1000 Maximum Function Calls. . . . . . . . . . . . . . 2147483647 Maximum CPU Time . . . . . . . . . . . . . . . . 604800 ABSGCONV Gradient Criterion . . . . . . . . . . . 0.0000100 GCONV Gradient Criterion . . . . . . . . . . . . 1E-8 ABSFCONV Function Criterion . . . . . . . . . . . 0 FCONV Function Criterion . . . . . . . . . . . . 0.0001000 FCONV2 Function Criterion . . . . . . . . . . . . 0 FSIZE Parameter . . . . . . . . . . . . . . . . . 0 ABSXCONV Parameter Change Criterion . . . . . . . 0 XCONV Parameter Change Criterion . . . . . . . . 0 XSIZE Parameter . . . . . . . . . . . . . . . . . 0 ABSCONV Function Criterion . . . . . . . . . . . -1.341E154 *** Other Control Parameters *** Line Search Method 2: Starting Alpha . . . . . . 1.00000 Line Search Precision LSPRECISION . . . . . . . . 0.40000 DAMPSTEP Parameter for Line Search . . . . . . . . Singularity Tolerance (SINGULAR) . . . . . . . . 1E-8 Hill & Plateau Data MLP with 30 Hidden Units Dual Quasi-Newton Optimization Dual Broyden - Fletcher - Goldfarb - Shanno Update (DBFGS) Number of Parameter Estimates 121 Optimization Start: Active Constraints= 0 Criterion= 0.030 Maximum Gradient Element= 0.006 Iter rest nfun act optcrit difcrit maxgrad alpha slope 1 0 4 0 0.0295 0.000041 0.00581 0.00225 -0.0366 2 0 6 0 0.0294 0.000099 0.0145 1.949 -0.0001 3 0 8 0 0.0286 0.000844 0.00638 2.096 -0.0008 4 0 9 0 0.0285 0.000034 0.0152 1.000 -0.0002 5 0 11 0 0.0281 0.000388 0.0119 3.029 -0.0003 6 0 13 0 0.0279 0.000244 0.00928 2.339 -0.0002 7 0 15 0 0.0274 0.000543 0.0259 8.913 -0.0001 8 0 16 0 0.0271 0.000249 0.0174 1.560 -0.0009 9 0 17 0 0.0269 0.000243 0.00931 1.422 -0.0004 10 0 18 0 0.0266 0.000247 0.0165 5.537 -0.0001 11 0 20 0 0.0252 0.00142 0.0294 7.860 -0.0004 12 0 22 0 0.0245 0.000745 0.0104 1.614 -0.0009 13 0 23 0 0.0241 0.000382 0.0367 5.977 -0.0003 14 0 24 0 0.0235 0.000613 0.0177 0.737 -0.0013 15 0 25 0 0.0233 0.00019 0.0322 2.407 -0.0006 16 0 27 0 0.0226 0.00064 0.00916 1.459 -0.0009 17 0 28 0 0.0216 0.00104 0.0128 4.153 -0.0004 18 0 30 0 0.0212 0.000426 0.00864 0.744 -0.0009 19 0 32 0 0.0209 0.000299 0.0107 2.066 -0.0003 20 0 34 0 0.0200 0.00082 0.0182 3.925 -0.0004 21 0 36 0 0.0195 0.000538 0.00766 1.034 -0.0011 22 0 38 0 0.0193 0.000188 0.00977 1.467 -0.0002 Yüklə 3,07 Mb. Dostları ilə paylaş:
|