2019/06/23 by Bradley J. Pillay, Absalom E. Ezugwu, Pillay, Bradley J. +1
Computer Science · Decision Sciences · Engineering · #Energy Load and Power Forecasting #FOS: Computer and information sciences #FOS: Economics and business #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Neural Networks and Applications #Statistical Finance (q-fin.ST) #Stock Market Forecasting Methods
paper · pdf · doi:10.48550/arxiv.1906.10121
openalex publication_date 2019/06/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
The prediction of stock prices is an important task in economics, investment\nand making financial decisions. This has, for decades, spurred the interest of\nmany researchers to make focused contributions to the design of accurate stock\nprice predictive models; of which some have been utilized to predict the next\nday opening and closing prices of the stock indices. This paper proposes the\ndesign and implementation of a hybrid symbiotic organisms search trained\nfeedforward neural network model for effective and accurate stock price\nprediction. The symbiotic organisms search algorithm is used as an efficient\noptimization technique to train the feedforward neural networks, while the\nresulting training process is used to build a better stock price prediction\nmodel. Furthermore, the study also presents a comparative performance\nevaluation of three different stock price forecasting models; namely, the\nparticle swarm optimization trained feedforward neural network model, the\ngenetic algorithm trained feedforward neural network model and the well-known\nARIMA model. The system developed in support of this study utilizes sixteen\nstock indices as time series datasets for training and testing purpose. Three\nstatistical evaluation measures are used to compare the results of the\nimplemented models, namely the root mean squared error, the mean absolute\npercentage error and the mean absolution deviation. The computational results\nobtained reveal that the symbiotic organisms search trained feedforward neural\nnetwork model exhibits outstanding predictive performance compared to the other\nmodels. However, the performance study shows that the three metaheuristics\ntrained feedforward neural network models have promising predictive competence\nfor solving problems of high dimensional nonlinear time series data, which are\ndifficult to capture by traditional models.\n