Text
PENERAPAN JARINGAN SYARAF TIRUAN UNTUK MEMPREDIKSI PERTUMBUHAN KARET (HEVEA BRASILIENSIS MUELL ARG) KLON BPM 24 DAN PB 260 PADA FASE PERTAMA
This study aims to test the artificial neural network model that was built to predict the growth of the rubber plant clones of BPM 24 and PB 260 clones in the first phase. The artificial neural network used in this study uses the backpropagation algorithm, using Matlab software. The artificial neural network architecture in this study uses 3 input layers, 1 output layer, 1 hidden layer, the activation function used in the study is logsig as a hidden layer and for the output layer using the purelin function. The rubber plant growth forecasting developed using artificial neural networks has a neuron structure of 16, 1 hidden layer, and the learning rate used is 0.1. training of artificial neural networks in forecasting the growth of rubber plants, the sample data used is 70 data from 105 data, and for network testing using 53 data samples from 105 data. After testing the network of each rubber clone, the BPM 24 regression value was R = 0.98711 and PB260 was R = 0.99379 and the mean square error (MSE) value obtained in the BPM 24 clone was 0.00092 on epoch 119 and the PB 260 clone was 0.00089 on epoch 73. Evaluation of the network model using the MAPE equation on the BPM 24 clone was quite good, with the MAPE value of shoot diameter and rootstock height less than 15%, but for rootstock diameter it was still not good with MAPE values above 15%. Evaluation on PB260 clone had greater MAPE value compared to BPM 24 clone with MAPE value of shoot diameter above 20%.
Inventory Code | Barcode | Call Number | Location | Status |
---|---|---|---|---|
2107002580 | T50923 | T509232021 | Central Library (REFERENCES) | Available but not for loan - Not for Loan |
No other version available