letters.chk

著名的神经网络工具箱

=============================================================================
	This file contains data to test the learning functions
=============================================================================

Network-Files:	letters_untrained.net
==============	

Pattern-Files:	letters.pat
==============	

Description:	- load network letters_untrained.net	
============	- load patterns letters.pat	
		- do NOT press init button
		- set cycles to 50
		- set shuffle off 
		- for each test set:
			- set learing function given in test set.
			- set learning parameters given in test set.
			- set update function given in test set.
			- set update parameters given in test set.
			- press the all button
			- compare the results with the given results 
			- reload network letters_untrained.net	

Test set 1)
===========

learning function  : Std_Backpropagation          0.2   0.0   0.0   0.0   0.0
update function    : Topological_Order		  0.0   0.0   0.0   0.0   0.0

  epoch:         SSE            MSE            SSE/o-units

     50:      93.51189        3.59661           3.59661
     45:      25.55399        0.98285           0.98285
     40:      24.93429        0.95901           0.95901
     35:      24.73275        0.95126           0.95126
     30:      24.59843        0.94609           0.94609
     25:      24.47465        0.94133           0.94133
     20:      24.34736        0.93644           0.93644
     15:      24.21373        0.93130           0.93130
     10:      24.07410        0.92593           0.92593
      5:      23.92981        0.92038           0.92038
      1:      23.81207        0.91585           0.91585


Test set 2)
===========

learning function  : Backpropagation with Momentum 0.8   0.6   0.1   0.0   0.0
update function    : Topological_Order		   0.0   0.0   0.0   0.0   0.0

  epoch:         SSE            MSE            SSE/o-units

     50:      37.47221        1.44124           1.44124
     45:      24.42921        0.93959           0.93959
     40:      20.02836        0.77032           0.77032
     35:      15.32081        0.58926           0.58926
     30:      10.93757        0.42068           0.42068
     25:       8.12217        0.31239           0.31239
     20:       5.43421        0.20901           0.20901
     15:       3.56318        0.13705           0.13705
     10:       1.95615        0.07524           0.07524
      5:       1.18959        0.04575           0.04575
      1:       0.90604        0.03485           0.03485


Test set 3)
===========

learning function  : Quickprop                     0.2   1.75  0.0001 0.0   0.0
update function    : Topological_Order		   0.0   0.0   0.0    0.0   0.0

  epoch:         SSE            MSE            SSE/o-units

     50:     165.45184        6.36353           6.36353
     45:      24.47066        0.94118           0.94118
     40:      19.36107        0.74466           0.74466
     35:      12.89132        0.49582           0.49582
     30:       8.85628        0.34063           0.34063
     25:       5.43688        0.20911           0.20911
     20:       3.31145        0.12736           0.12736
     15:       2.22553        0.08560           0.08560
     10:       1.85069        0.07118           0.07118
      5:       1.37056        0.05271           0.05271
      1:       1.97464        0.07595           0.07595


Test set 4)
===========

learning function  : Rprop                         0.2   0.0   0.0   0.0   0.0
update function    : Topological_Order		   0.0   0.0   0.0   0.0   0.0

  epoch:         SSE            MSE            SSE/o-units

     50:     165.45184        6.36353           6.36353
     45:      23.87504        0.91827           0.91827
     40:      17.03602        0.65523           0.65523
     35:      10.34510        0.39789           0.39789
     30:       5.06381        0.19476           0.19476
     25:       2.24841        0.08648           0.08648
     20:       1.55044        0.05963           0.05963
     15:       0.08736        0.00336           0.00336
     10:       0.00980        0.00038           0.00038
      5:       0.00075        0.00003           0.00003
      1:       0.00011        0.00000           0.00000