Change search
ReferencesLink to record
Permanent link

Direct link
Parameter Estimation of a non-Equilibrium Asset Pricing Model and Performance Analysis of the Calibration in Terms of Sloppiness
KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
2014 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

Prices of assets traded in stock markets often exhibit out of equilibrium behaviours, e.g. bubbles and recessions. Yukalov et al. have developed a model to describe these dynamics, and this Master thesis focuses on the problem of calibrating it using an Evolutionary algorithm and the Simulated Annealing method. In general, the parameter estimation performs far from desired, and a Sloppy model analysis of the deterministic system shows that the performance is linked to the sloppiness structure of the model. Accounting for sloppiness, the calibration results can be seen in a different light and the model could still be useful for predictions. Thus, the prediction performance on both synthetic and real-world data is studied, with good results in artificial markets and poor performance using real prices.

Abstract [sv]

Aktier handlas ofta för priser som skiljer sig från jämviktspriser, exempelvis under finansbubblor eller i recessioner. Yukalov et al. har tagit fram en modell för att beskriva dessa beteenden, och i den här Masteruppsatsen undersöks modellkalibrering genom en Evolutionär algoritm och ’the Simulated Annealing method’. Generellt är modellparametrarna dåligt uppskattade och en Sloppy model-analys av det deterministiska systemet visar att kalibreringsresultatet är beroende av modellens sloppiness struktur. Med detta i åtanke kan kalibreringsresultatet tolkas annorlunda och modellen kan fortfarande vara användbar för prediktion. Således är prediktionsprecisionen studerad för både syntestisk och riktig data, med god precision för simulerade marknader men sämre resultat för verkliga priser.

Place, publisher, year, edition, pages
TRITA-MAT-E, 2014:31
National Category
Mathematical Analysis
URN: urn:nbn:se:kth:diva-146275OAI: diva2:723414
Subject / course
Mathematical Statistics
Educational program
Master of Science - Mathematics
Available from: 2014-06-10 Created: 2014-06-10 Last updated: 2014-06-10Bibliographically approved

Open Access in DiVA

fulltext(1061 kB)70 downloads
File information
File name FULLTEXT01.pdfFile size 1061 kBChecksum SHA-512
Type fulltextMimetype application/pdf

By organisation
Mathematical Statistics
Mathematical Analysis

Search outside of DiVA

GoogleGoogle Scholar
Total: 70 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 150 hits
ReferencesLink to record
Permanent link

Direct link