scispace - formally typeset
Search or ask a question
Book ChapterDOI

Prediction of Stock Market Indices by Artificial Neural Networks Using Forecasting Algorithms

TL;DR: This model can guide any kind of a user with or without experience in the stock market to make profitable investments and is most likely to yield the closest prediction values with modest error rates.
Abstract: Application of artificially intelligent methods for predictions is a fairly old area, although it is also the one in which there is always room for improvement in performance and in consistency, given the escalating nature of information and the varying efficacy of prediction logics. A hybrid of simple statistical methods coupled with intelligent computing (here artificial neural networks) is most likely to yield the closest prediction values with modest error rates. We propose to build an analytical and predictive model for estimating the stock market indices. This model can guide any kind of a user with or without experience in the stock market to make profitable investments. The forecasting done is by way of three statistical algorithms and an adaptive, intelligent algorithm, thus making the process fairly robust. Training and testing the neural network will be done with two-month stock market index values for some of the companies listed with the Bombay Stock Exchange. A comparative result of the four algorithms is calculated, and the one with best precision is suggested to the user with a sale/buy/hold answer.
Citations
More filters
Journal ArticleDOI
TL;DR: In this article, a hybridized method that relies on using the support vector regression (SVR) method with equilibrium optimizer (EO) is proposed to foresee the closing prices of Egyptian Exchange (EGX).
Abstract: A hybridized method that relies on using the support vector regression (SVR) method with equilibrium optimizer (EO) is proposed to foresee the closing prices of Egyptian Exchange (EGX). Three indices are modeled and employed: EGX 30, EGX 30 capped, and EGX 50 EWI. The efficiency of using the technical indicators and statistical measures in the forecasting process is evaluated. The proposed EO-SVR-based forecasting model is adopted and evaluated using mean absolute percentage error, average, standard deviation, best fit, worst fit, and CPU time. Also, it is compared with recently developed metaheuristic optimization algorithms published in the literature such as whale optimization algorithm, salp swarm algorithm, Harris Hawks optimization, gray wolf optimizer, Henry gas solubility optimization, Barnacles mating optimizer, Manta ray foraging optimization, and slime mold algorithm. The proposed EO-SVR model got better results than other the counterparts, and EO-SVR is considered the optimal model according to its superior outcomes. Moreover, there is no need to use technical indicators and statistical measures as their effect is not noticeable.

21 citations

Journal ArticleDOI
TL;DR: This work explores the efficiency of three deep learning techniques, namely Bayesian regularization, Levenberg–Marquardt (lM), and scaled conjugate gradient (SCG), for training nonlinear autoregressive artificial neural networks (NARX) for predicting specifically the closing price of the Egyptian Stock Exchange indices.
Abstract: Financial analysis of the stock market using the historical data is the exigent demand in business and academia. This work explores the efficiency of three deep learning (Dl) techniques, namely Bayesian regularization (BE), Levenberg–Marquardt (lM), and scaled conjugate gradient (SCG), for training nonlinear autoregressive artificial neural networks (NARX) for predicting specifically the closing price of the Egyptian Stock Exchange indices (EGX-30, EGX-30-Capped, EGX-50-EWI, EGX-70, EGX-100, and NIlE). An empirical comparison is established among the experimented prediction models considering all techniques for the time horizon of 1 day, 3 days, 5 days, 7 days, 5 days and 30 days in advance, applying on all the datasets used in this study. For performance evaluation, statistical measures such as mean squared error (MSE) and correlation R are used. From the simulation result, it can be clearly suggested that BR outperforms other models for short-term prediction especially for 3 days ahead. On the other hand, lM generates better prediction accuracy than BR- and SCG-based models for long-term prediction, especially for 7-day prediction.

18 citations

Book ChapterDOI
06 Dec 2018
TL;DR: While linear models show better performance for BSE, artificial neural network based models exhibit higher predictive modeling accuracy for NSE, the design aspects are outlined for augmenting intelligent market prediction systems.
Abstract: The study aims to assess the major predictors of stock index closing using select set of technical and fundamental indicators from market data. Here two of major service sector specific indices of Bombay stock exchange (BSE) and National stock exchange (NSE) with historical data from 2004 up to 2016 are considered. By experimental simulation, the predictive estimates of index closing using automatic linear modeling, time-series based forecasting, and also artificial neural network models are analyzed. While linear models show better performance for BSE, artificial neural network based models exhibit higher predictive modeling accuracy for NSE. The design aspects are outlined for augmenting intelligent market prediction systems.

1 citations

References
More filters
Proceedings ArticleDOI
15 Apr 2013
TL;DR: This work investigates neural network models for time series in forecasting and it is found that the right parameters number of epochs, learning rate and momentum is 2960, 0.28 and 0.5 respectively for forecasting network by conducting various experiment.
Abstract: Forecasting based on time series data for stock prices, currency exchange rate, price indices, etc., is one of the active research areas in many field viz., finance, mathematics, physics, machine learning, etc. Initially, the problem of financial time sequences analysis and prediction are solved by many statistical models. During the past few decades, a large number of neural network models have been proposed to solve the problem of financial data and to obtain accurate prediction result. The statistical model integrated with ANN (Hybrid model) has given better result than using single model. This work discusses some basic ideas of time series data, need of ANN, importance of stock indices, survey of the previous works and it investigates neural network models for time series in forecasting. The forecasting accuracy is analyzed and measured with reference to an Indian stock market index such as Bombay Stock Exchange (BSE) and NIFTY MIDCAP50 in this study and it is found that the right parameters number of epochs, learning rate and momentum is 2960, 0.28 and 0.5 respectively for forecasting network by conducting various experiment.

66 citations

Journal ArticleDOI
TL;DR: The present study provided direct support for the potential use of accurate forecasts in decision making and fisheries management in the Mediterranean Sea by revealing a strong autoregressive character providing relatively high R2 and satisfactory forecasts that were close to the recorded CPUE values.
Abstract: Univariate and multivariate autoregressive integrated moving average (ARIMA) models were used to model and forecast the monthly pelagic production of fish species in the Mediterranean Sea during 1990–2005. Autocorrelation (AC) and partial autocorrelation (PAC) functions were estimated, which led to the identification and construction of seasonal ARIMA models, suitable in explaining the time series and forecasting the future catch per unit of effort (CPUE) values. Univariate and multivariate ARIMA models satisfactorily predicted the total pelagic fish production and the production of anchovy, sardine, and horse mackerel. The univariate ARIMA models demonstrated a good perpormance in terms of explained variability and predicting power. The current findings revealed a strong autoregressive character providing relatively high R 2 and satisfactory forecasts that were close to the recorded CPUE values. The present results also indicated that the multivariate ARIMA outperformed the univariate ARIMA models in terms of fitting accuracy. The opposite was evidenced when testing the forecasting accuracy of the two methods, where the univariate ARIMA models overall performed better than the multivariate models. The observed seasonal pattern in the monthly production series was attributed to the intrinsic nature of the pelagic fishery. As anchovy, sardine, and horse mackerel represent main target species in the Mediterranean pelagic fishery, the findings of the present study provided direct support for the potential use of accurate forecasts in decision making and fisheries management in the Mediterranean Sea.

57 citations

Proceedings ArticleDOI
26 Sep 2012
TL;DR: A hybrid system based on a Multi-Agent Architecture that will investigate the evolution of some neural network methods along with technical and fundamental analysis methods on stock market indexes and how this information influences the stock market behavior in order to improve the profitability of a short or medium time period investment.
Abstract: The goal of this paper is to create a hybrid system based on a Multi-Agent Architecture that will investigate the evolution of some neural network methods along with technical and fundamental analysis methods on stock market indexes and how this information influences the stock market behavior in order to improve the profitability of a short or medium time period investment. The proposed system compares the results of Standard Feed Forward Neural Network, Elman and Jordan Recurrent Neural Networks and a Neural Network evolved with Neuro Evolution of Augmenting Topologies (NEAT) in order to investigate which network gives the most accurate result and time performance by taking in consideration the close price of a stock. The system also finds correlations between the pattern recognition methods and technical and fundamental methods results in order to find the direction of the market trend, to predict the next day price of a stock and to trigger a useful buy/sell signal. We are also interested in finding a correlation between the evolution of price, volume, number of transactions in order to have a better view on which is the effect of stock liquidity on a stock price. In order to validate our model a prototype was developed and applied to the Bucharest Stock Exchange Market indexes.

29 citations

Proceedings ArticleDOI
01 Feb 2016
TL;DR: Experimental results show that the multi-binary classification using OAA technique outperforms other techniques and can provide the return on investment greater than the traditional analysis techniques.
Abstract: In stock market, successful investors can earn maximum profits depended on a stock selection and a suitable time on trading. Generally, investors use two statistical techniques for making a decision, which are the fundamental analysis and the technical analysis. Recently, machine learning models which are a part of artificial intelligence, has been applied to enhance investors for investment. A number of machine learning models have been investigated for stock prediction such as Genetic Algorithms (GAs), Support Vector Machines (SVMs) and Neural Network (NN). In this paper, several multiclass classification techniques using neural networks are investigated. The multi-binary classification experiments using One-Against-One (OAO) and One-Against-All (OAA) techniques are tested and they are compared with the traditional neural network. Furthermore, an alternative data preparation and a data selection process are proposed. The experimental results show that the multi-binary classification using OAA technique outperforms other techniques. It can provide the return on investment greater than the traditional analysis techniques.

24 citations

Journal ArticleDOI
TL;DR: A new methodology for the optimization of ANN parameters as it introduces a process of training ANN which is effective and less human-dependent and resulting in ANN where satisfactory performance is achieved.
Abstract: Determination of optimum feed forward artificial neural network (ANN) design and training parameters is an extremely important mission. It is a challenging and daunting task to find an ANN design, which is effective and accurate. This paper presents a new methodology for the optimization of ANN parameters as it introduces a process of training ANN which is effective and less human-dependent. The derived ANN achieves satisfactory performance and solves the time-consuming task of training process. A Genetic Algorithm (GA) has been used to optimize training algorithms, network architecture (i.e. number of hidden layer and neurons per layer), activation functions, initial weight, learning rate, momentum rate, and number of iterations. The preliminary result of the proposed approach has indicated that the new methodology can optimize designing and training parameters precisely, and resulting in ANN where satisfactory performance.

19 citations