Info

此问题已关闭。 请重新打开它进行编辑或回答。

why does training returns different output values with the target sets. using backpropagation

1 次查看(过去 30 天)
i am learning how to build text document image extraction and recognition using backpropagation.
but my ouput training seems wasn't like what i expected.
the output wasn't similiar with my target sets at all.
the target data are:
target = zeros(1,total_images);
target(:,[1 35 61 83 102]) = 'a';
target(:,[2 36 62 84 103]) = 'b';
target(:,[3 37 63 104]) = 'c';
target(:,[4 38 64 85 105]) = 'd';
target(:,[5 39 65 86 106]) = 'e';
target(:,[6 40 66 107]) = 'f';
target(:,[7 41 67 87 108]) = 'g';
target(:,[8 42 68 88 109]) = 'h';
target(:,[9 43 69 110]) = 'i';
target(:,127) = 'I';
target(:,[10 43 69 89 111]) = 'j';
target(:,[11 44 70 90 112]) = 'k';
target(:,[12 45 71 91 113]) = 'l';
target(:,[13 46 72 92 114]) = 'm';
target(:,[14 47 73 93 115]) = 'n';
target(:,[15 48 74 94 116]) = 'o';
target(:,[16 49 75 95 117]) = 'p';
target(:,[17 50 76 118]) = 'q';
target(:,[18 51 77 96 119]) = 'r';
target(:,[19 52 78 97 120]) = 's';
target(:,[20 53 79 98 121]) = 't';
target(:,21) = 'u';
target(:,55) = 'U';
target(:,[22 56 123]) = 'v';
target(:,[23 57 80 99 124]) = 'w';
target(:,[24 58 125]) = 'x';
target(:,[25 59 81 100 126]) = 'y';
target(:,[26 60 101]) = 'z';
target(:,37) = 0;
target(:,28) = 1;
target(:,29) = 2;
target(:,30) = 3;
target(:,31) = 4;
target(:,32) = 5;
target(:,33) = 6;
target(:,34) = 7;
but the result gives :
output =
Columns 1 through 22
96 95 96 113 117 95 98 98 94 107 107 98 112 112 116 114 93 92 116 111 111 95
Columns 23 through 44
100 110 108 110 115 104 96 106 110 101 102 113 117 96 93 110 113 111 95 98 96 94
Columns 45 through 66
107 112 98 112 112 111 110 112 92 93 103 105 100 93 108 116 96 93 110 113 111 111
Columns 67 through 88
98 107 94 107 107 98 112 112 111 110 112 92 116 103 116 108 102 93 113 111 98 98
Columns 89 through 110
107 95 106 117 112 111 94 115 93 103 95 111 117 115 96 93 110 113 111 111 98 107
Columns 111 through 127
94 107 107 98 112 112 111 110 112 92 116 103 116 100 98 108 115
instead of the target sets.
i've attached my training file (pelatihan_jst.m)
hopefully somebody would help me to solve this problem as soon as possible.
thanks before.

回答(0 个)

此问题已关闭。

产品


版本

R2015b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by