Teaching creativity through actuarial projects S t efanos Orfanos DePaul Universit y 2019 Act uarial Teaching Conference
Who needs creativity? Academic exercises: • Focus on speed • Apply a given method • Emphasize the final answer Real-life questions: • Have many acceptable answers • Require assumptions or interpretation, and dealing with ambiguity • Take time
How to teach creativity? Math-heavy programs: • Go over proofs of results • Have students come up with their own proofs Actuarial programs: • Assign proj ects that go beyond the theory • Introduce ambiguity • Impose realistic constraints on time and data
Examples Fit distributions to data. S tart with made-up data, then real loss data, and then big data. Be aware of the principle of parsimony . Work with data from various lines of business, e.g. personal or commercial auto, workers comp, medical malpractice, product liability, flood insurance, marine insurance… Apply misspecificat ion t est s and informat ion crit eria to formally evaluate the model fit. Consider the credibilit y of company loss data. Adj ust for spat ial or t emporal correlation. Develop and t rend loss data.
Suggestions Do: Provide guidance if students are confused Encourage experimentation and resourcefulness Reward continuous improvement Require clearly stated assumptions and sanity checks Don’ t: Prescribe what the answer should look like
Communicating creativity Primary concerns: • S pelling, grammar, syntax • Writing to be read Last but not least: • S tructure • Use of evocative graphs and visualizations • Recommendation
Recommend
More recommend
