A method of constructing a dynamic model of bankruptcies of economic entities

Author(s):  L.I. Kasimova, "Bashkir state University" branch Sterlitamak, Sterlitamak, Russia

A.N. Biryukov, Dr., associate Professor, "Bashkir state University" branch Sterlitamak, Sterlitamak, Russia, biryukov_str@mail.ru

Issue:  Volume 45, №1

Rubric:  Labor market and economic educablic and business

Annotation:  The object of the consideration of this work is very deep penetration requirements for an effective neural network learning algorithms in preprocessing processing. Developed a method to assess the adequacy of neural network models in the absence of any a priori information about the distribution law of noise in the data. This is the scientific novelty of this article, as this method allows interrelated to control the quality of preprocessing financial data processing and their quality of approximation in neural networks for assessing the risk of bankruptcy of an economic object. The main purpose of all algorithms preprocessing data processing is to improve the homogeneity of the data and improving their quality (information value) in the aspect of neural network learning. The result is the construction of a basic neural network dynamic recovery model, the multivariate "generalized production functions" on the basis of the Bayesian approach. The quality control neural network model is necessary for finding the projection horizon the risk of bankruptcy of the entity, the evaluation stage of the developing process of the crisis of the object in time. The effect from these theoretical proposals – improving the efficiency (in terms of compromise between prediction accuracy and robustness) neural network model for a complex simulation (strong noise in the data and lack of observations)

Keywords:  neural network model, the probability of the risk of bankruptcies, the blocks of the algorithm, accounting and reporting, managerial decision

Full text (PDF):  Download

Downloads count:  433