Measures Of Population Stability and Instability

Tom, Daniel (2022): Measures Of Population Stability and Instability.

Abstract

A popularly used PSI treats all variables as categorical, regardless of bin ordering. Also, bin boundaries and the number of bins could affect the PSI quantity. We build our PSI without requiring binning, distinguishes numeric shifts from categorical redistribution, and unify the two for mixed numeric/categorical variables.

• Item Type: MPRA Paper
• Language: English
• Keywords: Modeling, Time, Vintage, Shift, Redistribution, Stability, Instability, Mid-CDF, Jensen-Shannon, AABC, PSI, Numeric, Missing, Mixed, Ordinal, Categorical
• Subjects: C - Mathematical and Quantitative Methods > C6 - Mathematical Methods ; Programming Models ; Mathematical and Simulation Modeling > C65 - Miscellaneous Mathematical Tools

References

1. Kullback, S.; Leibler, R.A. (1951). "On information and sufficiency" Annals of Mathematical Statistics. 22.

2. Jeffreys, H. (1948). Theory of Probability, 2nd ed. The Clarendon Press, Oxford.

3. Lin, J. (1991). "Divergence measures based on the Shannon entropy" IEEE Transactions on Information Theory. 37 (1): 145–151.

4. Vaserstein, L. N. (1969). "Markov processes over denumerable products of spaces, describing large systems of automata" Problemy Peredači Informacii. 5 (3): 64–72.

5. Anderson, T. W. (1962). "On the Distribution of the Two-Sample Cramer–von Mises Criterion" Annals of Mathematical Statistics. Institute of Mathematical Statistics. 33 (3): 1148–1159.

6. Lancaster, H. O. (1961). "Significance tests in discrete distributions" Journal of the American Statistical Association, 56, 223-234.

7. Parzen, E. (2009). "United Applicable Statistics: Mid-Distribution, Mid-Quantile, Mid P Confidence Intervals Proportion p" University of Maryland.