Hi, when I applied the code from The Joker tutorial section 'Continue generating samples with standard MCMC' to a source with good periodicity, the rejection sampling in the first step of The Joker was able to produce a high-quality orbital solution.
Orbital parameters obtained from The Joker sampling: period (P) = 2.77381769 days, eccentricity (e) = 0.02842552, semi-amplitude (K) = 78.02635686 km/s, systemic velocity (v₀) = 17.84151884 km/s
Phase-folded diagram plotted using the sampling results.
However, the subsequent MCMC sampling results exhibited a widely-separated multi-modal distribution, making it impossible to plot a clean, Gaussian-like peak. I used 8 chains with 1000 sampling points per chain as the parameters for MCMC sampling.
Here is the posterior probability histogram.
However, when I generated the corner figure using only the chain with the highest posterior probability, it produced a relatively good result, even though this doesn't represent the true orbital parameters.
Here is the code I used.
https://github.com/SakakinoKonomi/project/blob/main/RunJoker.ipynb
Hi, when I applied the code from The Joker tutorial section 'Continue generating samples with standard MCMC' to a source with good periodicity, the rejection sampling in the first step of The Joker was able to produce a high-quality orbital solution.
Orbital parameters obtained from The Joker sampling: period (P) = 2.77381769 days, eccentricity (e) = 0.02842552, semi-amplitude (K) = 78.02635686 km/s, systemic velocity (v₀) = 17.84151884 km/s
Phase-folded diagram plotted using the sampling results.
However, the subsequent MCMC sampling results exhibited a widely-separated multi-modal distribution, making it impossible to plot a clean, Gaussian-like peak. I used 8 chains with 1000 sampling points per chain as the parameters for MCMC sampling.
Here is the posterior probability histogram.
However, when I generated the corner figure using only the chain with the highest posterior probability, it produced a relatively good result, even though this doesn't represent the true orbital parameters.
Here is the code I used.
https://github.com/SakakinoKonomi/project/blob/main/RunJoker.ipynb