Statistical Analysis and Prediction of Paddy Yield Using Neural Network

Abstract

Abstract. This research uses boxplot, Anova and posthoc to analyse the effect of factors such as urine and phosphorous in rice paddy yield. Then an artificial neural network (ANN) is used to predict paddy yields based on those factors. ANN is also used to predict paddy yields from polybags based on the actual data of paddy yields from rice field. A total of 25 data were used in this study where 70% data were used for training while 15% data each for testing and validation. We use the training model using data from rice field to predict paddy yield in polybags. STATISTICA software was used to run the neural networks. The predictive power of constructed neural networks was measured using accuracy measurement Mean Squared Error. The result shows that prediction can be made through neural network since the performance is very encouraging. Keywords: neural networks, paddy yield, prediction, statistical analysis.