Best Alpha for Forecasting Stock using Brown's Weighted Exponential Moving Average

Abstract

Stocks are a very profitable type of investment, one of the most profitable stock sectors is mining. PT. Aneka Tambang Tbk (ANTM) is the largest mining company in Indonesia engaged in metals and minerals. The problem that occurs in stock investment is the fluctuation of stock prices. Stock price fluctuations occur because of the process of buying and selling shares, rumors about the company, or certain policies made by the government that can affect the company. These rapid stock price fluctuations are very dangerous for investors because these fluctuations are not always directly proportional to the company's fundamental conditions. Therefore we need a system that can perform forecasting to monitor stock price fluctuations that occur. The author uses the Brown's Weighted Exponential Moving Average (B-WEMA) method for forecasting. The data used are 241 daily PT Aneka Tambang Tbk (ANTM) in the Covid-19 era, starting January 1, 2020 - December 31, 2020. The experiment was carried out into 3 scenarios, namely using 3 months, 6 months, and 12 months data. Alpha tested is 0.1 to 0.9 (1 digit behind the comma) while the number of moving averages is 3, 5, 10, 20, 50, 100, and 200 adjusted to the amount of data. From the experiment it was found that the best forecasting was obtained in scenario 2 with the best manual alpha value having a MAPE error value of 2.84% using an alpha of 0.9 and a moving average of 100.