For the data in the example on page 87 of the notes, we settled on a possible model corresponding to a distribution that factors as p(x1,x2,x3)=p(x1)p(x2|x1)p(x3|x2).
a) Find a undirected graph G such that the discrete DAG model M(G) is all discrete distributions having such a factorization.
b) Decide if M(G) is a good fit for the data-generating distribution based on the data on page 87 of the notes. Be sure to justify your decision with an (approximate) p-value.