Week 10-GSoC 2020

This blogs describes the 10th week of the program. Some of the highlights of this week are:

Merged PRs:

• Added sampling from Matrix Distributions
  • This PR adds the support of sampling for newly added matrix distributions from external libraries.
• Added sampling for Joint Random Variables
  • This PR closes one of the issue of sampling from Joint Distributions. Also, adding sampling support from few frequently used Joint Distributions.
• Change return type of P and E with evaluate=False
  • This PR changes the return type of P and E with evaluate=False. They are made to return their respective classes with evaluate=False instead of unevaluated Integral or Sum.
• Allow more than one RV in compound distribution
  • This PR changes the marginalisation algorithm of the Compound Distributions and allows to handle more than one Random Variables as parameters.
• Added Matrix Normal and Wishart Distribution
  • This PR adds Matrix Normal Distribution and Wishart Distributions and closes one of the stalled PR of Kumar Ritesh.

This week I am working to add the support for Mixture distributions. I would try to complete it by this week and start to work on the Assumptions of random variables.