File for storing a PND |
File for storing weights and parameters of a PND ensemble
The file stores a PND, including all the neural nets composing the ensemble.
The file includes the following information:
The suffix of the file is .poli.
- Version number and name of the file
- Number of classes, number of dichotomizers
- The decomposition matrix
- For each dichotomizer: number of layers, input, output, hidden layer dimensions, type of activation function, matrices of weights
Example:
Version: 1 d6.poli Number_class: 6 Number_dico: 6 Decomp_matrix: 6 6 1 -1 -1 -1 -1 -1 -1 1 -1 -1 -1 -1 -1 -1 1 -1 -1 -1 -1 -1 -1 1 -1 -1 -1 -1 -1 -1 1 -1 -1 -1 -1 -1 -1 1 Dichotomizer n. 1 Number_layers: 2 Dim_output: 1 Act.function: 1 Dim_hidden1: 7 Act.function: 1 Dim_input: 2 Weight matrix output layer : -0.84869243 +0.50773446 -1.20820605 -1.18508944 -0.50139631 +1.15460962 -1.97928685 Bias vector: +0.05229848 Weight matrix hidden layer : +0.79448978 -1.51513088 -0.41827703 +0.45628670 -2.12598336 -0.16895854 -0.62812573 -0.81626030 -1.05425689 -0.02124678 -0.06261009 +0.98244997 -0.82358696 -1.68786065 Bias vector: -0.07365713 +0.93976057 -0.11559361 +0.38876887 -0.13332475 +0.40650898 +0.35184726 Dichotomizer n. 2 Number_layers: +2 Dim_output: 1 Act.function: 1 Dim_hidden1: 7 Act.function: 1 Dim_input: 2 Weight matrix output layer : -1.55271915 -2.20561198 -0.58450233 -3.82153835 -0.47074284 +4.21468576 -2.27513366 Bias vector: +0.18316335 Weight matrix hidden layer : +1.80772941 +0.39684516 -2.90713897 +2.14354250 +0.30780607 -0.72876497 +3.83306553 +0.31786572 +0.85671194 +0.22462678 +5.41540033 -0.02333124 +0.57617067 -2.29696378 Bias vector: -1.25270545 -0.02137527 -0.78363063 -2.81113125 -0.15018892 +0.36699870 -1.74856460 Dichotomizer n. 3 Number_layers: +2 Dim_output: 1 Act.function: 1 Dim_hidden1: 7 Act.function: 1 Dim_input: 2 Weight matrix output layer : -2.00459710 -4.85427290 -2.32480945 +0.51642059 +0.35372826 +0.42834908 +1.51993279 Bias vector: +0.53956415 Weight matrix hidden layer : +3.63989384 +0.86063828 -4.41371291 +2.07582076 +3.84544034 -0.46214644 +0.43571708 -0.14369948 +0.31708927 -0.74742078 +0.65346961 +1.51737801 -2.99683840 -0.84874626 Bias vector: +0.20816412 -2.51791799 -0.42932538 -0.07324992 -0.96397062 -0.87823705 +0.13226981 Dichotomizer n. 4 Number_layers: +2 Dim_output: 1 Act.function: 1 Dim_hidden1: 7 Act.function: 1 Dim_input: 2 Weight matrix output layer : +2.24931058 -0.97797916 +1.03844975 -1.76183635 -1.43254291 -3.60279831 -1.61154382 Bias vector: +1.36536034 Weight matrix hidden layer : +1.18394053 -0.25060150 -0.12399663 +0.96730774 +0.52147306 +0.05112912 -1.70597481 -0.19337483 -1.61938214 -0.10247916 -2.38890435 +1.72006940 -0.58140195 +1.26654233 Bias vector: -0.43641352 -0.88314340 +1.45167876 +0.66417087 +0.67920579 +2.94945535 +0.52219946 Dichotomizer n. 5 Number_layers: +2 Dim_output: 1 Act.function: 1 Dim_hidden1: 7 Act.function: 1 Dim_input: 2 Weight matrix output layer : -0.68608817 +0.88831741 +0.00891729 +3.27444578 -2.83463495 -3.44310926 +1.49754612 Bias vector: -0.92065125 Weight matrix hidden layer : +0.44509417 +0.66074535 -0.91793219 +0.68053485 -0.45517193 +0.45061090 -2.52105774 -0.34344570 +2.64775526 -2.62318377 +0.44215540 +3.59420775 -0.62330258 -1.57580900 Bias vector: +0.46593660 +0.14182596 +0.50125235 -2.76217904 +0.46096204 +0.08451298 -0.55771326 Dichotomizer n. 6 Number_layers: +2 Dim_output: 1 Act.function: 1 Dim_hidden1: 7 Act.function: 1 Dim_input: 2 Weight matrix output layer : -2.24945905 -2.62949315 -1.37452983 +1.69397065 -0.94801265 -0.50938219 +2.33449689 Bias vector: -0.56995789 Weight matrix hidden layer : +1.59052594 -1.26942446 -0.77408837 -2.74130379 +0.62947456 -1.59425557 -1.90939876 -0.22444294 -0.00110161 -1.27750429 -0.23177998 -0.90854584 -2.36103919 -0.26706373 Bias vector: +1.09793796 -0.82328499 +0.55885538 +0.61709873 -1.00371385 -1.45164365 +0.66931998 END
Alphabetic index Hierarchy of classes