Pemodelan Sistem Prediksi Kebutuhan Energi pada Tanaman Kentang (Solanum tuberosum L.) dengan Metode Deep Learning

Abstract

Agriculture and energy have a very close relationship because the agricultural sector is one of the most important sectors that consumes and supplies energy. In this study, modeling was carried out to be able to predict the energy requirements of potato plants during the cultivation process from planting to harvest with SPSS and Jupyter notebook. The resulting Cobb-Douglas model is lnYi = 3,910 - 0,056lnX1 - 1,082lnX2 - 0,295lnX3 - 0,450lnX4 - 1,627lnX5 - 1,644lner and obtain an R² value of 1 or an accuracy of 100 %. Then built a system with deep learning methods, namely the Convolutional Neual Network (CNN) to classification of growth phases with potato plant images. The CNN model was designed on a Jupyter notebook with 100 epochs and a total of 5 layers and used 1125 image data of potato plants consisting of two classes, namely the vegetative phase with an energy requirement of 4195.80 MJ/Ha and the generative phase of 35746.45 MJ/ Ha. From this test it produces a classification with an accuracy value of 99%.