Prediksi Penggunaan Energi Listrik PLTMH Sengkaling I di RUSUNAWA UMM Menggunakan C4.5 dan Recurrent Neural Network (RNN)

ABSRACT Electrical energy consumption is required in almost all human activities, the need for electricity is quite dependent on public demand. Consumption patterns tend to increase from time to time, these changes are due to the increasingly modern human activities. To find out the pattern of electricity energy use, prediction and classification of each activity carried out are needed. This study aims to predict the use of electrical energy and classify variables that affect consumption in RUSUNAWA. Recurrent neural network is one of the methods used for time-series data and variable-shaped data can be processed using the C4.5 algorithm. The system used in this research uses the Recurrent Neural Network (RNN) method. To support the performance of the RNN method, the C4.5 algorithm is also analyzed and seen whether its performance is aligned. Based on the tests that have been carried out, RNN modeling produces a value of 0.01076 RMSE and 0.8966% MAPE, and C4.5 modeling produces accuracy of 93.44% precision of 95.45% and recall of 93.30%