Export is one of the most important ways for the country to generate income, which can have an impact on the country's economic stability. This research aims to forecast the value of Indonesian exports to G20 member countries. The Long Short-Term Memory method is used in this research to examine historical data on Indonesian exports from the previous 16 years. Experimental results show that the LSTM Neural Network method has the ability to predict the value of Indonesian exports to G20 member countries with a sufficient level of accuracy. The predictions generated by the model provide insight into trends and fluctuations in the value of exports in the future. The results of this study provide insight into the potential application of artificial intelligence techniques in economic and trade analysis. The results demonstrate that the LSTM model is capable of producing relatively accurate predictions, with an average score of Root Mean Square Error (RMSE) on training data is 0.10 and on testing data is 0.13, as well as graphs of prediction results demonstrating that the LSTM model can capture patterns and trends from Indonesia's export data to G20 countries. According to the prediction results, the highest export value to China is expected to be $6,100,000 in the 200th month (or in the year 2039), while the lowest export value to Mexico is expected to be $27,000 in the 135th month (or in the year 2034).
Copyrights © 2023