Potential Questions that could be asked at Conference

• choice of dataset can be a concern: does framework help with choice of dataset
• are stresses multiplicative or additive in relation to current level of yield curve
• can data be amended to allow for features or can this approach prevent that
• application towards time-evolving stochastic model Ans: no, pca does not model time evolution

bayesian method is for sparse noisy data but gilt yields abundant

regulation discourages novel approaches

We are looking at merits of model in its own right
things that may appeal to regulation are systematisation of expert judgements