Вестник МГТУ, 2023, Т. 26, № 4.

Пашенцев С. В. Нейронные сети как инструмент совершенствования математической модели. Рис. 17. Результаты выполнения программы Fig. 17. The results of running the program File Edit Train Statistics Run Options Window Help SSI П E«m llS||[;;bj| _ u i_i©j м м _ i* !i m m \ 1 M Ш ГJetwork Editor (TestModel_120, 01) Error function I Sum-squared 3 Layer Units Delete [ Width (i § Connections Shown |All layers * Act fn ILogistic » PSP fn ILinear ^ hi«01 Jhl«02 h1«03 hi«04 hi «05 h1#06 h1«07 h1«08 h2«01 h2«02 h2«03 VAR17 Threshold 0.217087 0.3210215 -0.5088 0.3461872 0.2123942 -1.133292 -0.1912 0.2217951 0.6224243 0.2922363 0.9694061 -0.5276 VAR1 0.5759804 -0.2139 -0.1851 0.5102043 -1.01189 0.3493258 -0.8352 0.1341888 VAR2 0.1457057 -0.1049 -0.691 -0.1351 0.4384726 -0.7048 -0.983 0.06327 VAR3 -0.€064 0. 00839 0.172269 0 .5174006 -0.0135 0.06254 -О. 5-54 0.3293989 VAR4 0.3437804 -0.3382 -0.1214 0 .002801 -0.03254 -0 .472927 0.7647526 0.2636704 VAR5 0.3594457 0.8335396 -0.4407 -0.716 0.2292609 -0.457787 -0.3916 -0.1588 VAR6 -0.7563 0.8094553 -0.2841 0.1111783 -0.1856 0.1562756 -0.9218 -0.1646 VAR7 0.4444219 0.8684042 1.176852 -0.458 0.4962717 0.2587392 0.2899545 0.2026335 VAR8 0.9477319 -0.07037 -0.3507 0.186558 -0.828381 0.8546913 -0.198 -0.583438 VAR9 0.3697394 -0.020375 0.1835338 -0.5462 -1. 007248 -0.1506 0 .4896804 0.01982 VAR10 -0.4158 -0.108 0.01938 0.4378547 -0.3112 0.08182 0.5784873 0.1377023 VAR11 -0.7546 0.6379606 0.2386985 0.4448303 -0.2077 0.106598 -0.2245 0.3523047 VAR12 -1.025825 0.9381955 -0.1789 -0.6865 -0.8869 -0.6691 -0.9415 0.24306 VAR13 -0.759158 0.9762368 0.6014601 -0.7881 -0.7151 -0.4337 0.9621863 0.2399583 VAR14 0.1231254 0.3579847 -0.1108 0.2200864 0.9131269 0.1760071 0.5991597 -0.1238 VAR15 -0.5021 -0.2101 0.9095391 0.09096 -0.1454 -1.253334 0.5167753 -1.901281 VAR16 0.7906255 0.5466644 -0.8027 0.6218805 0.8730862 -0.7761 0.6315441 0.5553287 hi «01 -0.02505 -0.7256 0.9500563 h1«02 0.243909 0.05702 0.2110641 hi «03 0.6188767 -0 .003659 0.9872862 hi «04 0.3196231 -0.8578 -0.3647 hi «05 -0.1848 -0.01648 -0.702573 hi «06 -0.6309 0.2766878 0.9203362 hi «07 -0.3352 -0.7559 -0.8489 hi «08 -0.701 0.8835748 0.9212403 h2«01 -0.6776 h2«02 0.1744832 h2«03 0.0599916 Рис. 18. Весовые матрицы и векторы смещения между слоями сети Fig. 18. Weight matrices and bias vectors between network layers 486