0.25
0.5
0.75
1.25
1.5
1.75
2
Path Integral Method for Estimation of Time Series
Published on Feb 25, 20078950 Views
Related categories
Chapter list
Data Assimilation00:00
Collaborators01:01
Three Estimation Problems01:15
Turning a model into a state estimation problem03:08
Statement of the Problem04:52
GOAL: estimate moments06:50
A Nonlinear Example07:25
Observations09:45
Extended Kalman Filter10:18
Alternative Approaches13:46
Observations15:37
KSP Filter Results15:40
Why not KSP?18:02
A Statistical-Mechanical Digression18:42
Fact: log n! ¼ n log n - n20:04
BAYESIAN STATEMENT20:30
Path Integral Method21:22
Path Integral Method0123:21
Path Integral Method0224:39
Path Integral Method0325:39
Otherwise use sampling27:08
Hybrid Monte Carlo27:41
The HMC algorithm29:42
What’s going on?33:07
What’s going on?0134:05
Unigrid Monte Carlo34:51
Generalized HMC35:19
PIMC Results37:14
RESULTS: decorrelation time38:06
Conclusions (Sampling)39:07
OVERALL CONCLUSIONS39:46
Further Information41:06