Jaringan Syaraf Tiruan untuk Pendugaan Porositas Tanah
| Main Authors: | Suharyatun, Siti, Rahmawati, W., Sugianti, C. |
|---|---|
| Other Authors: | Siti Suharyatun, Winda Rahmawati, Cicih Sugianti Universitas Lampung Jurusa Teknik Pertanian |
| Format: | Article info application/pdf Proceeding |
| Bahasa: | eng |
| Terbitan: |
Pusat Unggulan Riset Pengembangan Lahan Suboptimal (PUR-PLSO) Universitas Sriwijaya
, 2019
|
| Online Access: |
http://conference.unsri.ac.id/index.php/lahansuboptimal/article/view/1566 http://conference.unsri.ac.id/index.php/lahansuboptimal/article/view/1566/944 |
Daftar Isi:
- Suharyatun S, Rahmawati W, Sugianti C. 2019. Artificial neural networks for estimating soil porosity. In: Herlinda S et al. (Eds.), Prosiding Seminar Nasional Lahan Suboptimal 2019, Palembang 4-5 September 2019. pp. 424-429. Palembang: Unsri Press. Texture is one of the physical properties of soil that is permanent and related to the other physical properties of soil. Its properties is included the ability to absorb the water which expressed by porosity. This study aims to develop an artificial neural network (ANN) model to estimate soil porosity based on soil texture. The research was conducted in several stages, namely: (1) measuring the physical properties of soil consisting of texture, moisture content, volume density, and particle density (2) calculating soil porosity, (3) Developing ANN models are correlated on textures and soil porosity. ANN models are made using 3 input layer neurons, 5 hidden layer 1 neurons, 5 hidden layer 2 neurons and 1 input layer neurom. The ANN model uses the logig-tansig-purelin activation function with a RMSE (Root Mean Square Error) value of 2.0242. Determination of the training model (R2) was 0.957. The results of testing the validity of the models produce the same determination value R2 = 0.957. Keywords: artificial neural network, the