New Approach for E-Commerce Stock Prices Prediction : Combination of Machine Learning and Technical Analysis

Kelvin Lee, Yong Ming and Mohamad, Jais and Pick Soon, Lee (2022) New Approach for E-Commerce Stock Prices Prediction : Combination of Machine Learning and Technical Analysis. International Journal of Academic Research in Accounting Finance and Management Sciences,, 12 (3). pp. 653-665. ISSN 2222-6990

PDF New Approach.pdf Download (636kB)

Abstract

Forecasting stock market is always a challenge task for the investors. This study aimed to develop a new approach for forecasting the price movements of e-commerce stocks. The signals emitted by the technical indicators are used as the features for two machine learning algorithms in predicting the stocks movements. The technical indicators used in this study were Moving Average (MA), Moving Average Convergence Divergence (MACD), Relative Strength Index (RSI) and Stochastic Oscillator (SO). Meanwhile, the machine learning algorithms used in this study were Random Forest (RF) and K-Neighbor Nearest (KNN). The findings of this study indicated that the inclusion of signals emitted by MA rule with 5-days short MA and 20-days long MA helps to reduce the error values for the prediction model. Besides that, this study also found that the signals from MA, MACD, RSI and SO fit the prediction model well. The investors are recommended to use machine learning algorithms to predict the price movements of e-commerce stocks. Lastly, investors are recommended to consider the signals from these four technical indicators, MA (5-days short MA & 20 long-MA), MACD, RSI and SO as the reference for their investment strategies in e-commerce stocks.

Actions (For repository members only: login required)

View Item