| Popis: |
Moving Average (MA) has been developed by many researchers, economists, analysts, and other professionals into several techniques to give better results for time series analysis. Conventional methods, such as Simple Moving Average (SMA), Weighted Moving Average (WMA), and Exponential Moving Average (EMA), show its capability in predicting several kinds of time-series data by utilizing a web-based application. In some cases, conventional methods display significant error values in forecasting such kind of data including large capital stocks data in Indonesia Composite Index. Through previous research, it was then discovered that Holt's Double Exponential Smoothing (H-DES) formula can be combined with the WMA formula to be combined as a new formula that was later referred as Holt's Weighted Exponential Moving Average (H-WEMA). This H-WEMA was used to predict Indonesia Composite Index data and so was never used to predict data from individual large capital stock. With this research, the H-WEMA formula is tested to predict stocks data from large capital stocks market. The rate of error produced by H-WEMA calculated by MSE, MAPE, and MASE have results with error numbers getting smaller as the span data widens. All the stocks data that have the smallest error numbers are all on stock data that have some kind of trend in it. Using MSE, the calculated method resulted in three smallest error rates between price value of 2.17 to 2.20 as the insignificant movement in their range of price. Using MAPE, the smallest errors are found on three companies ranged from 1.19 to 1.34 percent while the top 3 biggest errors ranged from 6.7 to 6.9 percent. Using MASE, the three smallest errors are in the interval of 0.10 to 0.17 error value while the three largest errors produced are in the interval of 6.7 to 7.5 error value. |
