TY  - THES
N1  - Pembimbing :  Shofwatul ?Uyun, M.Kom

ID  - digilib12197
UR  - https://digilib.uin-suka.ac.id/id/eprint/12197/
A1  - RIZKI TUNJUNG SARI , NIM. 09650006
Y1  - 2013/07/08/
N2  - Research on the maturity of tomatoes recognition has been done by
various methods and produce different accuracy. Method with a high degree of
accuracy are implemented into a system. Therefore, a prototype is needed to
compare several methods in order to obtain a good degree of accuracy.
To identify the maturity of granola tomatoes, 63 tomatoes is captured
using digital camera and stored in computer and in .jpg extension sized 680x680.
The images are extracted the RGB values that becomes the main feature for the
maturity of granola tomatoes identification. Identification is using two methods,
Backpropagation Neural Network (BPNN) and Matching Algorithm for
comparing the results of the accuracy of both methods.
Matching algorithm that used is City Block, Euclidean, order 3 and 4 of
Minkowski. Using 63 images, it shows the results of testing the accuracy of City
Block, Euclidean, order 3 and 4 of Minkowski are 99,2 % and Euclidean has
smallest distance than others. Testing with BPNN shows that the result is also
99,2 %. Optimal architecture is input with 3 nodes, 1 hidden layer with 3 nodes,
and output layer with 2 nodes. The optimal parameters is MSE of 0, epoch of
1000, learning rate of 0,01, and mc of 0,6. But order 4 of Minkowski has the
smallest running time than BPNN then order 4 of Minkowski becomes the best
method to indentify the maturity of granola tomatoes.


PB  - UIN SUNAN KALIJAGA
KW  - Keywords : Backpropagation
KW  -  Matching Algorithm
KW  -  Granola Tomatoes




M1  - skripsi
TI  - PERBANDINGAN METODE JARINGAN SYARAF TIRUAN
BACKPROPAGATION DAN ALGORITMA PENCOCOKAN DALAM
MENGIDENTIFIKASI KEMATANGAN TOMAT BUAH
BERDASARKAN CIRI WARNA RGB


AV  - restricted
ER  -