Ramadhona, Gandhi (2018) Prediksi Produktivitas Padi Menggunakan Jaringan Syaraf Tiruan Backpropagation. Sarjana thesis, Universitas Brawijaya.
Abstract
Beras sangat penting bagi manusia, terutama masyarakat ASEAN. Salah satu negara yang membudidayakan beras adalah Indonesia. Pada tahun 2015, Indonesia merupakan negara ASEAN yang menduduki peringkat ketiga tertinggi dalam hal penghasil beras terbesar di dunia. Meski demikian Indonesia masih perlu mengimpor beras tiap tahunnya karena tingginya permintaan dan konsumsi perkapita masyarakat Indonesia. Selain itu perbedaan hasil panen ditiap daerah mengakibatkan kelangkaan beras karena tidak optimalnya teknik pertanian yang digunakan. Pada penelitian ini mengimplementasikan metode jaringan syaraf tiruan backpropagation untuk meramalkan hasil produktivitas padi. Dalam implementasinya, data dinormalisasi menggunakan min – max normalization dan inisialisasi bobot menggunakan Nguyen – Widrow. Berdasarkan hasil pengujian parameter untuk metode backpropagation, hasil RMSE yang paling minimum yakni 8.6918 dengan nilai parameter learning rate = 0.8, hidden layer = 3, hidden neuron = 4 dengan jumlah epoch 10000 terhadap 135 data latih dan 13 data uji. Berdasarkan hasil pengujian 5 fold cross validation terhadap kestabilan pengujian data mendapatkan nilai rata – rata RMSE sebesar 8.2126.
English Abstract
Rice is very important for human beings, especially to the ASEAN community. Indonesia is one of the ASEAN countries that cultivate rice. In 2015, Indonesia ranked as the third-highest in terms of the world's largest rice producer. However, Indonesia still have to import rice every year due to its high demand and to fulfil Indonesian's per-capita consumption. The other reason is the different amount of harvest on each areas resulting in a scarcity of rice because the country can not be able to optimize the farming techniques that are used. This research use the methods of backpropagation neural network to predict the results of the rice productivity. In its implementation, the data is normalized using the min – max normalization and weighting initialization using Nguyen – Widrow. Based on the results of testing the parameters for the method of backpropagation, shows the most minimum RMSE i.e. 8.6918 with parameter values learning rate = 0.8, hidden layer neurons, hidden = 3 = 4 with the number of epoch 10000 against 135 training and 13 test data. Based on result of 5 fold cross validation against the stability testing data gets an average RMSE of 8.2126.
| Item Type: | Thesis (Sarjana) |
|---|---|
| Identification Number: | SKR/FTIK/2018/473/051808078 |
| Uncontrolled Keywords: | prediksi, produktivitas padi, jaringan syaraf tiruan, backpropagation. predicting, rice productivity, articial neural network, backpropagation. |
| Subjects: | 000 Computer science, information and general works > 006 Special computer methods > 006.3 Artificial intelligence > 006.32 Neural nets (neural networks) |
| Divisions: | Fakultas Ilmu Komputer > Teknik Informatika |
| Depositing User: | Budi Wahyono Wahyono |
| Date Deposited: | 31 Jan 2019 02:47 |
| Last Modified: | 21 Oct 2021 01:48 |
| URI: | http://repository.ub.ac.id/id/eprint/13842 |
Preview |
Text
Gandhi Ramadhona.pdf Download (3MB) | Preview |
Actions (login required)
 |
View Item |