On-line change-point detection procedures for Initial Public Offerings
2010 (English)Independent thesis Advanced level (degree of Master (One Year)), 15 credits / 22,5 HE credits
Student thesis
Abstract [en]
In this thesis we investigate the case of monitoring of stocks havingjust been introduced for public trading on the nancial market. Theempirical distribution of the change-point for 20 assets for 60 days was calculated to check the support for the assumption that the priceinitially drop or rise to some steady level.The price process X = {Xt : t in Z} is assumed to be an AR(1) process with a shift in the mean value from a slope to a constant. The Shiryaev-Roberts, Shewhart, EWMA, Likelihood ratio and CUSUM proceduresfor detecting a change-point in such a process are derived. The expecteddelay of the motivated alarm according to these methods is achievedunder the assumptions of a Poisson, uniform, binomial and geometric distributed
by means of simulations.
Place, publisher, year, edition, pages
2010. , p. 55
Keywords [en]
Financial Mathematics, cange-point detection, monitoring
National Category
Computational Mathematics Probability Theory and Statistics
Identifiers
URN: urn:nbn:se:hh:diva-13940OAI: oai:DiVA.org:hh-13940DiVA, id: diva2:373130
Presentation
2010-06-04, Haldasalen, Halmstad University, Halmstad, 10:40 (English)
Uppsok
Physics, Chemistry, Mathematics
Supervisors
Examiners
2010-12-022010-11-302010-12-02Bibliographically approved