Author(s)

Ruiqing Xu

Corresponding Author

Ruiqing Xu

ABSTRACT

Due to the tremendous interests of several large corporations, making stock market predictions has always been one of the most activated research fields. However, forecasting stock prices has been restricted by huge factors regarding its characteristics of volatility, seasonality and unpredictability. In 1991, Grant Mcqueen and Steven Thorley, who are Professors of Finance at BYU Marriott School of Management, conduct examination of predicting stock prices by using Markov Chains. In 2005, Hassan and Nath used only one HMM that is trained on the past dataset of the chosen airlines to forecast the stock prices of these companies. After reading two journals regarding different approaches, it was not hard to notice that these methods actually were conducted based on contradicted assumptions. It appears that the transition of methods used in order to predict stock prices has been well developed, and this recent study in 2005 has overturned the past premise on random walks theory, which suggested stock prices can also be predicted by using existing patterns. Finally, people can improve the prediction accuracy by combining AI technology and HMM according to Hassan and Nath, which is the future researches in this field. After reading these all five journals, this paper focus on the trading history on the stock market of Alibaba Group. Then researcher sees the history data during the last five years as training data, so it is possible to make reasonable predictions for the following days based on the transition matrix, which is founded from the training data.

KEYWORDS

Stock prices, Patterns, Markova chain, Alibaba group, growth rate, transition Matrix