Prediksi Pendapatan Kargo Menggunakan Arsitektur Long Short Term Memory

Indonesian air cargo transportation is currently experiencing quite significant development. One of the cargo services in Indonesia is Garuda Indonesia Cargo and has several branch offices. The existence of a revenue forecast model is expected to provide insight into a branch office. This research proposes an income prediction using the Deep Learning algorithm, Long Short Term Memory (LSTM). LSTM is used because the data to be processed is time series data. The results of testing accuracy are measured using Root Mean Squared Error (RMSE). The data in this study are income from one branch office, the Cargo Service Center (CSC) Tangerang City. Data contains collections of goods delivery transactions every day. The data goes through 4 preprocessing processes, namely subtotal, outlier detection, difference, and scaling. The results of this study show the best prediction results, namely the composition of the 90% train data and 10% test data with RMSE values of train data 641375.70 and test data 594197.70