Stock prediction and trading using RVFL networks

• Main Author: Mehra, Manav
• Other Authors: Wang Lipo
• Format: Final Year Project
• Language: English
• Published: 2018
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/74732
• Institution: Nanyang Technological University

The ability to predict future changes in the stock market is a very powerful tool. Various machine learning techniques are being applied to try and make this task possible. This paper takes on a new approach to predicting stock prices making use of a Random Vector Functional Link (RVFL) network. Various RVFL networks were trained using MATLAB and their performances were compared to create a robust prediction model. The networks were trained and tested using closing values of the NASDAQ stock index. Two types of RVFL networks were created, one having a delay of 4 days and the other having a delay of 9 days. The RVFL network clearly performs better than other Artificial Neural Network (ANN) models in terms of both training times and prediction accuracy. Possible future directions are pointed out which could enable the creation of an RVFL model to be used in real world stock trading applications.