Improving Markov Chain Monte Carlo algorithms in LISA Pathfinder Data Analysis
暂无分享,去创建一个
The LISA Pathfinder mission (LPF) aims to test key technologies for the future LISA mission. The LISA Technology Package (LTP) on-board LPF will consist of an exhaustive suite of experiments and its outcome will be crucial for the future detection of gravitational waves. In order to achieve maximum sensitivity, we need to have an understanding of every instrument on-board and parametrize the properties of the underlying noise models. The Data Analysis team has developed algorithms for parameter estimation of the system. A very promising one implemented for LISA Pathfinder data analysis is the Markov Chain Monte Carlo. A series of experiments are going to take place during flight operations and each experiment is going to provide us with essential information for the next in the sequence. Therefore, it is a priority to optimize and improve our tools available for data analysis during the mission. Using a Bayesian framework analysis allows us to apply prior knowledge for each experiment, which means that we can efficiently use our prior estimates for the parameters, making the method more accurate and significantly faster. This, together with other algorithm improvements, will lead us to our main goal, which is no other than creating a robust and reliable tool for parameter estimation during the LPF mission.
[1] Bayesian parameter estimation in the second LISA Pathfinder mock data challenge , 2010, 1008.5280.
[2] Walter Fichter,et al. Data analysis for the LISA Technology Package , 2009, Classical and Quantum Gravity.
[3] David J. C. MacKay,et al. Information Theory, Inference, and Learning Algorithms , 2004, IEEE Transactions on Information Theory.
[4] Karsten Danzmann,et al. LISA - An ESA Cornerstone Mission for the Detection and Observation of Gravitational Waves , 2003 .