I. Introduction
Forecasting price trends accurately in the stock market is a formidable task, yet it is an economic reality. Despite the many caveats associated with financial forecasting, estimating values of financial indices is an important component of business decision-making [1]. There exist a wide variety of forecasting approaches, ranging from simple data analysis (e.g., moving average-based methods), to more complex techniques such as genetic algorithms [2]. These techniques can be categorized into three main classes [1, p. 88]:
Subjective forecasting, which is performed based on experience, intuition and guesswork. It is usually inferred from both macroeconomic (e.g., inflation rate) and microeconomic factors (e.g., yield rate of a particular stock).
Extrapolation techniques, whose aim is to project past trends into the future. Common extrapolation techniques include regression analysis and methods based on error criteria, such as the mean absolute deviation (MAD) and the mean squared error (MSE), to name a few.
Causal modeling, where the goal is to predict a lagging variable based on a leading variable. The relationship between these two variables can be modeled as a “cause” and “effect,” and is typically inferred from Bayesian methods. In this work, we use phase synchrony to model this relationship.
An algorithmic gold standard currently used in trading is the so-called Technical Analysis (TA), where based on the history of trading (e.g., price changes, volume of transactions) in a certain stock or in “the averages,” probable future trend is deduced [3]. Technical analysis indicators include relative strength index, stochastics, moving average-based methods, convergence/divergence, momentum oscillator, and commodity channel index. A commonly used technical indicator in trading is the moving average cross-over, calculated from two moving average processes of different time lengths (usually the 5− and 20-day moving average pairs). Whenever the slower 20-day moving average trace crosses below the faster 5-day moving average, this is considered a “buy” indicator (signal); conversely, whenever the 20-day moving average trace crosses above the 5-day moving average this is a “sell” indicator.