Search

This explanatory study investigates the existence of a predictive financial machine that can forecast future stock prices within the confines of acceptable accuracy and reliability, as it unravels the dynamics surrounding the prime objective of investors to maximize returns, based on technical rather than on fundamental analysis. The quest traverse from the tradition domain of econometrics, whose knowledge based has now incorporated many fields and techniques of bioscience (genetic algorithm and programing) and computer science (artificial neural networks, deep learning, fuzzy based), enabling investigation into concepts of prudential behavioral finance and quantum physics, into the emerging field of FinTech. The financial machine, at the disposal of stakeholders, is potentially an innovative legal public electronic money printing press, which, with an acceptable accuracy predicts future trends and pricing, taking the concept of profit maximization to a new dimension. While the concept of perfect prediction is not computationally possible, enlightened investors can benefit from superior gains, considering future stock market direction, and returns, on a daily, weekly and monthly basis, with different levels of confidence. Stakeholders may need to employ hybrid models, systematically and logically breaking tasks into stages, and selecting from an array of finovative techniques, keeping focus on the overall objective; machine learning techniques can extract latent knowledge from timeseries macroeconomic indicators, and even from crowd-sourced knowledge bases, to considerately enhance the quality of stock market forecasting. As FinTech develops into an evolving domain, yet to reach the maturity stage, the contemporary landscape comprises of, support vector machine (SVM), long short-term memory (LSTM), neural network, and machine learning primarily using US data, as the prominent financial stock market predicting techniques, although emerging markets also offer promising prospects. Nevertheless, in the contemporary stock market forecasting arena, the most common techniques employed are based on HMM, ANN, RNN and fuzzy-based deep learning, although given the complexities, and uncontrollable factors, the progressive finovative evolution is likely to continue unabated, in conformance with the adaptive market hypothesis (AMH). [ABSTRACT FROM AUTHOR]