ANALISIS TIME SERIES PADA DATA METEOROLOGI BMKG KOTA BANTEN MENGGUNAKAN METODE GATED RECURRENT UNIT

Abstract

In everyday life, weather and climate conditions greatly affect all human activities. Complete and accurate weather forecasts are needed to improve performance in various fields of human activity. Methods for making forecasts have been tested in various research processes, one of which is the Gated Recurrent Unit (GRU). The research was conducted using weather data in the form of average temperature, average humidity, and average wind speed with the Gated Recurrent Unit (GRU) method. The parameter that influenced the results of the GRU model prediction in this study was the epoch. The best average accuracy was obtained with an epoch of 100, and the best RMSE and MAE values were 1.31 and 1.04%.