About unstable MemFit

[李冠俊]

Hello,

I am Chun from Taiwan. Recently, I am doing data analysis for a continuous report task (colorwheel task) and try to use MemToolbox.

However, I always got poor fitting results (e.g. very exaggerated sd estimation) and the results fluctuate enormously whenever I redo analysis.

Even if I attempt to analyze the data 'errors' provided in the Demo1 in the tutorial, I can never get a reasonable result. 

(I just simply analyzed by errors by MemFit(errors) as in Demo1)

So, I am wondering that is this a normal phenomenon because of using MCMC algorithm for posterior inference

or is there any possible ways for me to get the fitting more stable.

note. I have checked my data format is correct.

Thanks, 

Chun