Yashu Seth

@jamieforth What is the status of this PR?

@Harshvardhan1 Appveyor has not been configured yet. Hence, it will fail for now. You need to add tests for your change. I think tests have not been added that is...

Hello everyone. Hope you are doing well. Where are we with this currently?

@ankurankan If the users have to provide an instance of the distribution to ```ContinuousFactor``` then why not use the individual distributions directly? If a particular method is unavailable in that...

@ankurankan This seems convenient. But it would require a number of if-else conditions in ```ContinuousFactor```. @khalibartan Which module uses a similar implementation in scikit-learn?

@ankurankan @khalibartan Cool. With this we can allow only ```ContinuousFactor``` to be associated with pgmpy.models for continuous cases.

@ankurankan I think we should begin work on this. A lot of other things depend on this.

@ankurankan I think we can rename the current ContinuousFactor class to BaseDistribution, such that it acts as a parent class for classes like JointGaussianDistribution, CanonicalFactor etc. Then have a separate...

@ankurankan Any update on this? I am free these days so if you want I can take this up after discussing the exact details.

> I think we can rename the current ContinuousFactor class to BaseDistribution, such that it acts as a parent class for classes like JointGaussianDistribution, CanonicalFactor etc. Then have a separate...