抽象的な

A Note on Exploring Discrete Choice Probability in the Light of Distorted Probability Estimates

Sibnarayan Guria and Pooja Sengupta

In this paper we provide a fresh look to the problem of exploration of discrete choice probability, drawing ideas from prospect theory. The deviation of the true choice probability estimates can be attributed to model misspecication and also to the personal choice preference of the individual decision makers. We look at the expressions of choice probability, and explore these in the context of prospect theory and risk aversion of a decision maker. The expression of the distortion of the choice probability with respect to both the above mentioned sources is discussed using the linear in log odds transformation. We then propose that in addition to maximising expected utility of a particular choice, a decision maker should choose a policy that will maximise the estimating power of the true choice probability as a linear ction of the distorted probability. This has been postulated in the form of a theorem, which has been established empirically in the last section of this paper. The current study shows encouraging results for exploiting the nuances of choice probability.