International trade is an important aspect in the progress of development because it acts as a driving force in the growth of the nation’s economy. In a macroeconomic perspective, export activities are the most important variable in determining a country’s economic system. As one of the developing countries, Indonesia has opened itself to take part in export activities to strengthen its economic position in the global market. This make export an international trade activity that plays an important role in the economic progress in Indonesia. In export activities, the export unit value index or export price index is an important component. This reflects the price development of export commodity groups and plays a significant role in assessing Indonesia’s economic performance. The Export Unit Value Index is specifically used to calculate Gross Domestic Product (GDP) at fixed prices and calculate the terms of trade.
One of Indonesia’s leading commodities that dominate the export market is jewelry. In export activities, the export unit value index is an important component that serves to describe the development of export commodity prices. This unit value index always changes every time and fluctuates. During January June 2021, Indonesia’s jewelry industry exports reached USD 1.23 billion. The figure almost doubled to USD 2.37 in January-June 2022. The large contribution of the jewelry industry sector to Indonesia’s export value was also seen during the Covid-19 pandemic in 2020. While other sectors experienced a decline, the jewelry industry sector actually increased by 76%. Jewelry commodities are categorized under the SITC code 897, which includes jewelry, gold and silver, and other articles made from precious or semi-precious materials. The commodity of the jewelry industry sector has a large contribution to Indonesia’s economic growth, but the condition of the Unit Value Index cannot be ascertained and can change at any time, therefore, a prediction is needed to analyze the value of the Unit Value Index for the next few periods.
This research conducts a comparative analysis of the performance of parametric method, non-parametric method, and machine learning, specifically, ARIMA, Fourier series estimator, and Support Vector Regression (SVR). This study aims to evaluate the effectiveness of various methods in improving prediction accuracy for the unit value index of the SITC code 897 in Indonesia. The research data used is secondary data including monthly export unit value index data with SITC code 897 in Indonesia obtained from the Central Bureau of Statistics. The data divided into 90% training data and 10% testing data. The methods used in this analysis are ARIMA, Fourier series estimator, and SVR. The best model obtained from each method is ARIMA (1,1,1) with MAPE of 10.92%, Fourier series estimator with MAPE of 8.47%, and an SVR RBF kernel function with MAPE of 3.73%. The results of this study obtained the best method for predicting the unit value index of SITC code 897 is SVR with an RMSE value of 8.288 and very good prediction accuracy.
Author: Elly Pusporani, S.Si., M.Stat.
