Financial Forecasting using Neural Networks under Multithreaded Environment

摘要

Neural Networks (NN) have been applied in all walks of day-to-day life, from credit report to mortgage/loan approval process from sorting and searching algorithms to traveling salesman problem. Parallel algorithms for some of these NN applications have been reported in the literature and training times are shown to be reduced to some extent. However, parallel algorithm for financial forecasting has not found its way due to multi-disciplinary involvement such as: Financial Economics, Neural Networks and Parallel Computing. The present paper attempts to solve the financial forecasting problem efficiently by developing a multithreaded algorithm using NN. We have developed a multithreaded algorithm and implemented it with a proposed NN architecture. The preliminary results are encouraging.