- 1 名前：1 [04/09/11 01:48:08]
- について教えてよ
- 227 名前：デフォルトの名無しさん mailto:sage [2007/11/15(木) 22:54:57 ]
- サンプルデータは次のようになってます。一つの行が1セットで、最初の15が入力、最後の1つが出力（教師信号）です。
0.743,0.500,0.455,0.727,0.455,0.818,0.636,0.091,0.455,0.636,0.818,0.318,0.500,0.045,0.318,0.700
0.486,0.000,0.682,0.682,0.136,0.500,0.591,0.136,0.409,0.500,0.682,0.545,0.591,0.409,0.318,0.500
0.714,0.469,0.636,0.545,0.545,0.455,0.364,0.545,0.636,0.636,0.545,0.364,0.545,0.727,0.545,0.600
0.714,0.594,0.136,0.727,0.636,0.909,0.727,0.455,0.727,0.455,0.727,0.409,0.227,0.045,0.227,0.200
0.829,0.781,0.727,0.636,0.727,0.545,0.636,0.727,0.727,0.818,0.727,0.591,0.591,0.500,0.682,0.400
0.771,0.313,0.636,0.727,0.364,0.727,0.909,0.000,0.364,0.818,1.000,0.318,0.591,0.045,0.227,0.700
0.771,0.531,0.000,0.818,0.545,1.000,0.818,1.000,0.636,0.636,0.909,0.500,0.773,0.227,0.318,0.400
0.429,0.281,1.000,0.727,0.545,0.455,0.727,0.636,0.818,0.727,0.727,0.591,0.318,0.864,0.682,0.500
0.771,0.313,0.545,0.545,0.909,0.455,0.455,0.636,0.636,0.545,0.545,0.773,0.500,0.318,0.318,0.500
0.486,0.438,0.727,0.909,0.545,0.545,0.455,0.727,0.727,0.636,0.818,0.500,0.500,0.227,0.500,0.300
0.857,0.563,0.000,0.909,0.000,0.455,0.818,0.000,0.273,0.636,0.727,0.318,0.318,0.227,0.136,0.600
0.714,0.375,0.455,0.727,0.727,0.818,0.455,0.455,0.545,0.273,0.000,0.045,0.682,0.045,0.136,0.300
0.200,0.844,0.182,0.909,0.000,0.545,0.636,0.636,0.273,0.545,0.727,0.955,0.409,0.318,0.682,0.200
0.800,0.469,0.364,0.636,0.727,0.636,0.727,0.545,0.545,0.636,0.909,0.864,0.136,0.136,0.500,0.600
0.800,0.281,0.636,0.455,0.818,0.909,0.545,0.727,0.909,0.273,0.364,0.455,0.818,0.545,0.273,0.600
0.829,0.438,0.364,0.545,0.000,0.727,0.545,0.455,0.636,0.545,0.455,0.227,0.773,0.227,0.227,0.500
0.714,0.500,0.182,0.818,0.000,1.000,0.636,0.273,0.545,0.727,0.545,0.227,0.955,0.318,0.318,0.600
0.600,0.344,0.909,0.818,0.273,0.909,0.455,0.182,0.455,0.545,0.818,0.227,0.682,0.409,0.409,0.300
0.629,0.500,0.273,0.818,0.000,0.818,0.818,0.273,0.364,0.727,0.455,0.864,0.500,0.409,0.227,0.300
- 228 名前：デフォルトの名無しさん mailto:sage [2007/11/15(木) 22:57:59 ]
- 続き
0.657,0.500,0.818,0.727,0.455,0.636,0.182,0.182,0.091,0.455,0.636,0.682,0.409,0.136,0.682,0.200
0.829,0.813,0.091,0.636,0.000,0.545,0.455,0.636,0.545,0.545,0.727,0.773,0.591,0.773,0.591,0.500
0.629,0.391,0.818,0.818,0.273,0.727,0.727,0.364,0.364,0.455,0.545,0.773,0.591,0.318,0.682,0.500
0.800,0.094,0.909,0.636,1.000,0.636,0.727,0.545,0.182,0.455,0.455,0.227,0.682,0.591,0.227,0.400
0.629,0.391,0.818,0.727,0.727,0.909,0.909,0.545,0.727,0.455,0.818,0.409,0.864,0.591,0.318,0.300
0.857,0.656,0.727,0.545,0.636,0.455,0.364,0.818,0.636,0.273,0.636,0.773,0.500,0.318,0.409,0.500
0.829,0.438,0.455,0.545,0.818,0.727,0.364,0.636,0.545,0.727,0.545,0.318,0.864,0.227,0.591,0.000
0.657,0.594,0.727,0.455,0.545,0.818,0.636,0.727,0.636,0.636,0.727,0.591,0.682,0.682,0.682,0.400
0.914,0.438,0.364,0.364,0.273,0.727,0.818,0.455,0.455,0.455,0.545,0.318,0.591,0.227,0.227,0.500
0.829,0.656,0.636,0.636,0.455,1.000,0.545,0.364,0.545,0.636,0.818,0.409,0.591,0.045,0.227,0.500
0.314,0.281,0.091,0.727,0.364,0.909,0.545,0.364,0.364,0.545,0.636,0.500,0.773,0.045,0.227,0.500
0.686,0.219,0.455,0.455,0.727,0.591,0.500,0.227,0.591,0.591,0.591,0.455,0.727,0.182,0.364,0.000
0.629,0.281,0.636,1.000,0.909,0.455,0.818,0.182,0.455,0.364,0.727,0.227,0.500,0.045,0.318,0.300
以上です。
プログラムは、熊沢逸夫著「の学習とニューラルネットワーク」という本のバックプロバゲーション部分をJavaで書き直した物で、
普通のデータ(XORを試しました)であれば、正常に学習、予測することは確認しています。
|
 |
