Support math methods and techniques for 0.1.0 RC

[iduartgomez]

0.1.0 will require addition of several mathematical techniques to support the core library:

• [x] Discrete bayesian networks marginals sampling.
• [x] Continuous bayesian networks marginals sampling.
• [x] Hybrid bayesian networks marginals sampling.
• [ ] (hybrid) Markov random fields marginal sampling.

Maybe add efficient conditioning schemes for constraint markov random fields as a bonus (see: Computing Marginal Distributions over Continuous Markov Networks for Statistical Relational Learning. Brocheler, Getoor. 2010)