FASCOM-stock : using a fuzzy associative conjunctive map in option trading

• Main Author: Chua, Wei Kar.
• Other Authors: Quek Hiok Chai
• Format: Final Year Project
• Language: English
• Published: 2013
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
• Online Access: http://hdl.handle.net/10356/55025
• Institution: Nanyang Technological University

Cortical maps are found in many biological and artificial neural systems. These maps organize and represent the information obtain from sensory inputs and play important roles in learning and memory process. In this project, the structure of a novel fuzzy neural network architecture FASCOM, short for Fuzzy Associative Cortical Maps, is being studied. The following architecture uses features inspired by the structure and functions of cortical maps. It also integrated a linguistic fuzzy model to perform associative learning if input-output pairs. However it lacks the ability to create appropriate number of fuzzy rules and does not have online learning. Various clustering techniques such as global k-means algorithm and discrete incremental clustering has been studied to enhance FASCOM in fuzzy rules creation. FASCOM is then applied to wide variety of classification problems. Then analysis on the performance of FASCOM using these clustering techniques is done. The relation of stock and options is also being analyzed by understanding of the nature of options and the mechanisms in options trading. An option trading system has been implemented with FASCOM as a prediction tool. Experiments have been conducted on to test the accuracy of FASCOM and the option trading system.