Artificial Neural Networks to Predict Melon (cucumis melo L.) Production in Tropical Greenhouse, Indonesia

Abstract

Quality of melon indicated by size (fruit weight), appearance, and sweetness. In Indonesia, weight of high quality melon was 800 to 1,200 grams each. Mainly, the melon was cultivated in open fields during the dry season with several limitations of cultivation. To cope with those problems, melon was cultivated inside the greenhouse. However, there are several parameters influenced to melon quality inside the tropical greenhouse with hydroponic system. There were a few studies on the prediction model development of melon inside the greenhouse in a tropical area, Indonesia. The aim of this study was to develop an artificial neural networks (ANNs) model to predict the melon production inside the greenhouse (fruit weight) using several parameters such as the number of days to fruit formation, number of days to maturity, plant length, fruit width, fruit length, fruit cavity diameter, flesh diameter, branch number, fruit branch number, and leaf number. The result of this study was the ANN model with configurations of 10 input layers, 6 hidden layers, and 1 output layer with R2 was 0.93. This study concluded that there is a correlation between the input parameters with the weight of the melon.