Email Updates RSS
Search

Neural Networks for Stock Price Prediction

By IGI Global on Apr 4, 2011
Stock values fluctuate on a daily basis, and accurate predictions regarding the direction and force of these changes are essential to making a profit in this volatile market. But what if you could accurately predict what the value of a particular stock would be one day in the future? Neural networks, programming which mimics the processes of the human brain, offer one way to make an accurate prediction.
"Though neural networks have been studied since 1940s they are relatively new methods for modeling and forecasting financial data, for example, stock/asset price," write three professors in their article published the 1st Quarter issue of IGI Global's International Journal of Grid and High Performance Computing (IJGHPC). "Kimoto et al. (1990) introduced modular neural networks that could learn the relationships between the past technical and economic indices and predict the timing to buy and sell a stock," they continue. "Simulation of buying and selling stocks using the prediction system showed an excellent profit. Stock price fluctuation factors could be extracted by analyzing the neural networks."


"In a sense neural networks are computer programs that are trained to learn analogously to a human brain remembering some information from the past," they assert.

However, training a neural network utilizing Backpropogation currently takes a substantial amount of time, with a training time of 60 hours "not uncommon" under "certain parametric conditions," write the three professors, Rashedur M. Rahman, of the North South University in Bangladesh, Ruppa K. Thulasiram of the University of Manitoba, Canada, and Parimala Thulasiraman, of the University of Manitoba, Canada. "This huge amount of training time is sometimes unacceptable for quick decision making that is generally required in financial markets," they write.
Professors Rahman, Thulasiram and Thulasiraman designed two "parallel and multithreaded neural network algorithms" in an attempt to improve this training speed. The first, "neuron parallelism" (NP), divides and distributes the neurons among a selected number of processors, with each processor working in tandem to make calculations. In the second method, "training set parallelism" (TSP), the data is divided among a select number of processors with the neural network reproduced on each.
Professors Rahman, Thulasiram and Thulasiraman found that their methods did speed the neural network's training process but they "did not achieve a linear speedup…" due to the resource-expensive process of synchronizing the parallel processors.
"There are opportunities to improve the performance with other types of algorithms such as pipelining the network," they write. "For example, in pipelining, the different weight layers are computed in different processing elements."
You can find more details about the International Journal of Grid and High Performance Computing, published quarterly, on our website at www.igi-global.com/ijghpc.

IGI Global publishes a multitude of research on Artificial Intelligence and Computer Engineering. Those interested in these topics or software development might be interested several of our forthcoming titles:

Read a free sample copy of the International Journal of Grid and High Performance Computing.

Browse for more posts in:
SoftwareScience & EngineeringBusiness Intelligence

No comments Comments

Log in or sign up to comment.
Be the first to comment!

More from IGI Global

Apr 18, 2024 Uniting Nations with Multinational Electronic Health Records to Stop Lassa Fever's Pandemic Threat!
The World Health Organization (WHO) has reported a surge in Lassa fever cases in Nigeria, emphasizing the urgent need for containment measures.
IGI GlobalRead More
Medical, Healthcare, and Life SciencesHealthcare Information SystemsBooks & E-BooksResearch Trends
Apr 16, 2024 IGI Global Announces the 2024 Journal Reviewers Award Winners
IGI Global congratulates the winners of this year's Journal Reviewer Award
IGI GlobalRead More
JournalsAwards & RecognitionOpen Access
Apr 15, 2024 What Defines Authorship and Who Should Be Included in a Manuscript’s Byline?
For decades, academic publishing has been plagued with discrepancies surrounding authorship of scholarly research...
IGI GlobalRead More
Books & E-BooksAcquisitions
Apr 15, 2024 Doody’s Selects IGI Global Books for Special Topics List
Two IGI Global publications have been recognized by Doody's for their excellence and niche topic focus.
IGI GlobalRead More
Medical, Healthcare, and Life SciencesMedia and CommunicationsBooks & E-BooksAwards & Recognition
Apr 11, 2024 Technology is Making Disparities Worse
Digital Inclusion Week underscored the urgent need for a national digital equity plan in the US due to disparities in internet access and digital skills.
IGI GlobalRead More
Apr 8, 2024 Nearly 90% of IGI Global’s Frontlist Books Now Indexed in Scopus
The majority of IGI Global's books Frontlist is now indexed by Scopus. Learn what this prestigious recognition means for the publisher and the experts behind these books.
IGI GlobalRead More
Books & E-BooksAwards & RecognitionReviews & Indexing
Apr 4, 2024 Navigating Neglected Maintenance and Safety Risks of Flight Travel
A United Airlines Flight 837 bound for Tokyo from San Francisco could not take off due to a mechanical issue.
IGI GlobalRead More
Business and ManagementComputer Science and Information TechnologyScience & EngineeringHospitality, Travel & Tourism ManagementAerospace EngineeringBooks & E-BooksResearch Trends
Apr 3, 2024 Introducing the New IGI Book Topic Brochures
IGI Global is excited to introduce a new initiative to provide detailed insights into various fascinating subjects. Our brochures offer a comprehensive overview of timely publications covering the latest research topics...
IGI GlobalRead More
EducationBusiness and ManagementComputer Science and Information TechnologySocial Sciences and HumanitiesBooks & E-Books
Apr 3, 2024 Interview with a Global Guru in Education
Hear from Dr. Velliaris, who was voted as a Top 30 Global Guru in Education.
EducationBooks & E-BooksInterviewAuthor News
Mar 29, 2024 Cocoa Prices Skyrocket
Cocoa prices surging to $10,000 per metric ton challenge chocolate manufacturers and consumers, risking supply chain stability and higher retail prices...
IGI GlobalRead More
Business and ManagementEnvironmental, Agricultural, and Physical SciencesSustainable DevelopmentBooks & E-BooksResearch Trends
First Previous 1 2 3 4 5 6 7 8 9 10  ... Next Last