Stock Market Prediction Performance of Neural Networks: A Literature Review

Abstract

In this paper, previous studies featuring an artificial neural networks based prediction model have been reviewed. The main purpose of this review is to examine studies which use directional prediction accuracy (also known as hit ratio) or profitability of the model as a benchmark since other forecast error measures -namely mean absolute deviation (MAD), root mean squared error (RMSE), mean absolute error (MAE) and mean squared error (MSE) -have been criticized for the argument that they are not able to actually show how useful the prediction model is, in terms of financial gains (i.e. for practical usage). In order to meet the publication selection criteria mentioned above, a large number of publications have been examined and 25 of papers satisfying the criteria are selected for comparison. Classification of the eligible papers are summarized in a table format for future studies.

M&A Short-Term Performance Based on Elman Neural Network Model: Evidence from 2006 to 2019 in China

“…Moreover, it has proven to be more suitable for stock forecasting than traditional linear models [56]. Ican and Çelik [57] compared 25 literatures based on neural network predicting stock prices, holding that selecting the appropriate stock data (input information) and neural network structure have an important influence on the fitting effect. According to the topology of neuron connections, neural networks can be divided into forward networks (such as BP neural networks) and feedback networks (such as Elman neural networks).…”

Decision Support System for Stock Portfolio Selection Using Artificial Intelligence and Machine Learning

“…There are many studies available in the literature that discuss the comparison of various AI/ML techniques in terms of their inputs and output prediction accuracy, even though these studies have varying degree of prediction accuracies, one common theme is that the accuracy and reliability of the model increases when fundamental financial variables [17] are used for the establishing the relationship between input variables and stock price [18]. Literature available in the field of AI/ML applications for stock price prediction are based on the popular ML models such as Artificial Neural Networks (ANN), Support Vector Machines (SVM) and Decision Tree techniques such as Model Trees and Random Forests have technical indicators as inputs [19]. There are studies on comparison of various ML techniques used for stock price prediction and have indicated that linear regression models cannot capture the complex non-linearity in the stock prices and have also indicated that ANNs along with Decision trees have better performance when fundamental financial variables are chosen as the model inputs [20].…”

Predicting the uptrend and downtrend in stock markets for intraday trading using machine learning algorithms

“…In [13] the economic time series has been analysed and it results in the similarity between a genuine time series and a series where one of the systematic elements is weak. In [15] the need for investors to predict the stock market is mentioned, and how this prediction helps retain investors' attention for stocks [14] the paper suggests that an amalgamation of artificial neural networks and any other statistical tool or machine learning algorithm provides better results for financial time series predictions. In [16] the research paper keeps into account factors like pricing, social behaviour, regulations and how it affects the stock market.…”