A1197
Title: Forecast of the expected number of decisions over time assuming two opposing decision-making patterns
Authors: Pawel Kopciuszewski - ING Hubs Poland (Poland) [presenting]
Elzbieta Kopciuszewska - Vistula University (Poland)
Abstract: A concept for analyzing how the expected number and variance of decisions made by two opposing decision-making patterns with different premises fluctuate over time is presented. Although the concept is general, it is presented intuitively using an example of two opposite groups of investors in a market. An investor from the first population makes a decision at time k only if the normalized difference between the population mean decision level and the investor's expectations for that level at time k exceeds a threshold specific to that investor. The second population follows the same statistical approach as the first but operates with opposite rationale for buying or selling the product. A formula is proven that shows how the expected values and variance of the number of decisions made over time depend on the initial statistical parameters of the population. Based on several simulations, results show how the expected value and variance of the fraction of products sold and bought change over time and when cyclicality can be observed. Several scenarios are considered, including cases where an investor's expectations about the decision level in the market at time k result from observations at time k-1 or from the trend observed from time k-2 to k-1.
