Forecasting marketplace stock value in Indonesia based on the best time series analysis model

Economic growth is one of the targets of the Sustainable Development Goals (SDGs) that the Indonesian government wants to achieve. It can be done by stabilize the stock market in the marketplace business. Bukalapak and Matahari are examples of online and offline-based marketplaces that most widely known by the Indonesian people. Therefore, this study aims to predict the market share value of Bukalapak and Matahari Department Store through various time series analysis methods. The methods used range from simple methods such as Autoregressive Integrated Moving Average (ARIMA), ARIMA Autoregressive Conditional Heteroscedasticity (ARCH) and Generalized Autoregressive Conditional Heteroscedasticity (GARCH) methods if the data has heteroscedasticity, to newly developed time series analysis, such as time series analysis with nonparametric regression using kernel and Fourier series estimator. The data used in this study is daily data on Bukalapak and Matahari stocks as much as 276 data which is divided into 90% in sample data and 10% out sample data. The best model obtained is nonparametric regression using kernel estimator with Gaussian function based on the smallest Generalized Cross Validation (GCV) and Akaike Information Criterion (AIC) values. The model is used as the basis for forecasting with MAPE which results in 6.7% and 2.7% for Bukalapak and Matahari stock data, respectively. These results indicate that the resulting model is good. The forecasting results can be used as recommendations and evaluations for both the government and economic activity actors so that they can prepare economic plans in order to achieve economic improvement targets in Indonesia.