<![CDATA[Rice is a crucial crop in Indonesia as it serves as a staple food. However, rice production is frequently hindered by pests, particularly the brown planthopper (BPH), which poses a serious threat to agricultural productivity in Gianyar District. To minimize crop failures and enhance productivity, predicting the spread of BPH on rice plants is crucial. In this study, a time series dataset consisting of 120 data points on BPH distribution from 2012 to 2021 was utilized. The data was split into 90% training data and 10% testing data. By employing the Recurrent Neural Network (RNN) architecture, the best-performing model achieved a minimal Root Mean Square Error (RMSE) value of 10.0503, with 500 epochs, a learning rate of 0.007, 5 neurons in the input layer, and 80 neurons in the hidden layer. This model also achieved a Mean Absolute Percentage Error (MAPE) value of 16.64%, indicating good predictive performance. The predictive results can be used by laboratories as decision support for rice productivity improvement strategies]]>
Copyrights © 2022
