Binary choice modelling based on the stable distribution
В. С. Тимофеев, Anastasiia A. Sanina · 2016
The paper considers the classification problem solving based on the binary choice models. Since the classification quality of the well known models depends on the nature of input and output variables it is supposed to build a new model based on any other universal distribution. It is the stable distribution. Moreover, an additional parameter optimization procedure for the selected distribution is viewed. It should be noted that this procedure affects the classification quality distinctly.