Yes it does seem like a bit of a 'trend' towards a greater focus on the mathematical sections of the course... but not much data to go on!
Credit migration and Euler seem emminently testable.
PCA and SVD is hard to test mathematically without the question dominating the paper - due to the time consuming calculations required. More likely perhaps, they could ask about: the potential applications; a description of the method; the advantages/disadvantages (including reducing dimensionality).
Not sure why Time Series hasn't been tested directly. Perhaps they regard it as having been largely covered by earlier subjects? However, unlike PCA and SVD they could realistically ask calculation questions. In addition they could test understanding of: model choice (e.g. ARMA vs GARCH); model output (interpreting correlograms, QQ plots etc.)
Last edited: Apr 27, 2015