At the moment we have the option to use CPnest or EMCEE. These are used to sample the prior distribution in KDE and to sample the posterior in Asy_peakbag.
Together with the addition of dimensionality reduction, I propose to implement Dynesty which has been tested extensively in relation to the dimensionality work and the performance is quite good.
It does require the PPF of the prior probability density, whereas CPnest and EMCEE work directly off the PDF.
So we need to make a decision about whether to keep the KDE approach and move completely to the PCA method for implementing the prior information, or we put in a switch between the two modes.
At the moment we have the option to use CPnest or EMCEE. These are used to sample the prior distribution in KDE and to sample the posterior in Asy_peakbag.
Together with the addition of dimensionality reduction, I propose to implement Dynesty which has been tested extensively in relation to the dimensionality work and the performance is quite good.
It does require the PPF of the prior probability density, whereas CPnest and EMCEE work directly off the PDF.
So we need to make a decision about whether to keep the KDE approach and move completely to the PCA method for implementing the prior information, or we put in a switch between the two modes.