An Empirical Study for Dynamic TIPP Policy Using XCS with Knowledge Rules
Mei-Chih Chen, Ming-Chia Huang, An-Pin Chen
Available Online October 2006.
- https://doi.org/10.2991/jcis.2006.123How to use a DOI?
- Multi-Agent System, Time Invariant Portfolio Protection, Extended Classifier System
- The purpose of this empirical study is intended to investigate XCS (Extended Classifier System) based model with knowledge rules for dynamic TIPP (Time Invariant Portfolio Protection) policy. There are two XCS-based agents in the proposed model (MA-TIPP).One agent dynamically optimizes Multiple and Tolerance variables which are concerned as the important parameters of TIPP and recommend trading. The other one is aimed to use 80% accuracy historical rules retained by classifier system to improve the previous agent prediction accuracy. The Multiple and Toleranc parameters which are optimized by GA and stock technical indexes such as Moving Average(MA), Moving Average Convergence and Divergence (MACD), Stochastic Line(KD), Relative Strength Index (RSI), Close and Volume are used as the input factors of classifier system. This proposed model is evaluated by 80% insurance and periods of TAIEX (TSEC weighted index) from 1996 to 2004. The experimental results are also compared with single XCS agent model (SA-TIPP) without using historical knowledge rules.
- Open Access
- This is an open access article distributed under the CC BY-NC license.
Cite this article
TY - CONF AU - Mei-Chih Chen AU - Ming-Chia Huang AU - An-Pin Chen PY - 2006/10 DA - 2006/10 TI - An Empirical Study for Dynamic TIPP Policy Using XCS with Knowledge Rules BT - 9th Joint International Conference on Information Sciences (JCIS-06) PB - Atlantis Press SN - 1951-6851 UR - https://doi.org/10.2991/jcis.2006.123 DO - https://doi.org/10.2991/jcis.2006.123 ID - Chen2006/10 ER -