Mortality cohort effect detection and measurement based on differential geometry

This paper analyzes mortality cohort effect of birth year and develops an approach to identify and measure cohort effects in mortality data set. The approach is based on differential geometry and leads to an explicit result which can describe how strong the cohort effect is in any period. This quantitative measurement provides a possibility to compare cohort effects among different countries or groups. The paper also suggests to use coefficient of variation as a measurement of the whole cohort effect for one country. Data of several countries are taken as examples to explain our approach and results.