CSPA case studies
CSPA · Put your knowledge to work in context.
The Certified Specialist in Predictive Analytics (CSPA) is an iCAS credential involving multiple learning and assessment components, practical analytical work, and professional responsibilities. It should not be treated as one uniform multiple-choice exam, and it is not interchangeable with the separate CAS PCPA assessment. The CSPA case studies are 6 cases and about 1,333 words.
1Reconciling two conflicting loss-cost reports
An insurer prepares a simplified loss-cost report for a clearly defined annual portfolio. The policy table contains one thousand earned policy-years.
226 words, about 1 min2A renewal dataset with late features and duplicated policies
A team builds a renewal prediction table scored at the beginning of each calendar year. The source contains policy-period rows, claim records, and multiple payment transactions per claim.
227 words, about 1 min3Interpreting a count model and a logistic review model
An insurer fits a Poisson count model with a log link and log exposure as a specified unit-coefficient offset. At a defined reference feature pattern, the fitted annual claim rate is 0.12.
216 words, about 1 min4Selecting and explaining a more flexible model
A team compares a small tree, a random forest, and a gradient-boosted model for the same defined insurance target. The deepest tree nearly eliminates training error but performs poorly on a later validation period.
217 words, about 1 min5A fraud ranking improvement with a worse review queue
A fraud-review team evaluates a model on a properly dated holdout sample. At one threshold the confusion matrix has thirty true positives, ten false positives, twenty false negatives, and forty true negatives.
217 words, about 1 min6A reproducible reserve project with responsible communication
An analyst prepares an insurance project using a model of payments during the next twelve months. The executive summary calls the output an estimate of all ultimate unpaid loss, although no additional development component is included.
230 words, about 1 min