Structural evolving appetitive reward-based pseudo-outer-product fuzzy neural network SE-ARPOP-CRI(S)

• Main Author: Do, The Anh
• Other Authors: Quek Hiok Chai
• Format: Final Year Project
• Language: English
• Published: 2012
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
• Online Access: http://hdl.handle.net/10356/48898
• Institution: Nanyang Technological University

Inspired by ARPOP-CRI(S) (Appetitive Reward-based Pseudo-Outer-Product Fuzzy Neural Network), a sequential learning model that incorporates the concepts of pre-synaptic and synaptic inspirations in Aplysia feeding behaviour, the student decided to build a new computational model that supports both structure learning as well as parameter learning in an online learning process. The model are constructed with preservation of features that support ARPOP-CRI(S) to deal with di culties in processing dynamic data stream as well as novel points that help the model with a better generality and performance. The model is evaluated and compared with several established works. Experimental results from benchmark applications support the model with promising results. The model is observed working well in a number of elds which include time-varying signal detection, forecasting, nonlinear control and de-cision support. The model is also tested extensively in real nancial market data. This report is the documentation representing steps and research taken as well as results recorded by the student in the course of accomplishing the project.