Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/38556
Title: Genetic complementary learning fuzzy neural network based on approximate analogical reasoning schema for stock market trend prediction
Authors: Lai, Jianxin
Keywords: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
Issue Date: 2010
Abstract: Over the past decade, there have been many attempts made to predict stock market data using statistical and data-mining models. While they all achieve a certain degree of success, they have certain major drawbacks, namely requiring long training times, results being difficult to understand and certain inconsistency with lack of accuracy in the predictions. All these drawbacks could result in loss of millions of dollars. Hence, it is paramount that the prediction results are as accurate as possible. Therefore, in this report, a Genetic Complementary Learning Fuzzy Neural Network based on Approximate Analogical Reasoning Schema (GCLFNN-AARS) is proposed to tackle the problem of long training times and poor accuracy. This system makes use of Genetic Algorithm (GA)’s capability to obtain optimal solution, the human-like recognition skills of hippocampal complementary learning and the Approximate Analogical Reasoning Schema (AARS)’s conceptual clarity in the hope of achieving better results. The system aims to avoid computational complexity with the use of AARS in the fuzzy inference process instead of the commonly-used Compositional Rule of Inference (CRI). The experimental results of the system show that it has the potential to be a useful tool for stock market trend prediction.
URI: http://hdl.handle.net/10356/38556
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:SCSE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
SCE09-0196.pdf
  Restricted Access
995.66 kBAdobe PDFView/Open

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.