A Multivariate Compound Poisson Model with Copula and its Application on Limit Order Book

• Main Authors: KUO,LING-YU, 郭玲佑
• Other Authors: HUANG,SHIH-FENG
• Format: Others
• Language: en_US
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/d39h98

碩士 === 國立高雄大學 === 統計學研究所 === 106 === A multivariate compound Poisson model with copula is proposed to depict the dynamics of Limit Order Book (LOB), where the intensity rates of order arrivals are assumed to have autocorrelations. The distribution of the first-passage time when the best ask (or bid) price moves is derived under the proposed model. A method of short-term stock price prediction is also developed. The LOB data of Intel, Microsoft, Johnson & Johnson, Yahoo and Taiwan Semiconductor Manufacturing stocks on five different days during 2008 and 2016 are employed for the empirical study. The numerical results indicate that the proposed model has satisfactory performances on modeling LOB and predicting short-term stock prices.