TY  - JOUR
TI  - Using neural networks for solvency prediction for russian companies of manufacturing industries
T2  - 
IS  - 
KW  - neural networks
KW  - solvency prediction
KW  - bankruptcy predicting models
KW  - logistic regression
AB  - Tatyana Bogdanova - Associate Professor, Department of Business Analytics, Faculty of Business Informatics, National Research University Higher School of Economics.Address: 20, Myasnitskaya str., Moscow, 101000, Russian Federation.E-mail: tanbog@hse.ruTimofey Shevgunov - Associate Professor, Department of Theoretical Radio Engineering, Moscow Aviation Institute.Address: 4, Volokolamskoe shosse, Moscow, 125993, Russian Federation.E-mail: shevgunov@gmail.comOlga Uvarova- Senior Lecturer, Department of Business Analytics, Faculty of Business Informatics, National Research University Higher School of Economics.    Address: 20, Myasnitskaya str., Moscow, 101000, Russian Federation.E-mail: ouvarova@hse.ruSelection of reliable business partners represents a typical task, where the decision-makers have to assess the financial position of a great number of potential contractors.The study comes up with a method of assessing long-term financial credibility of an enterprise based on processing of a system of financial performance indicators using neural networks. Neural networks are usually used in financial performance assessment applications for creating black-box models through predetermined algorithms and selected software. This can be suitable for a typical business user but keeps critical details of the subject area hidden from researchers and analysts. This study intends to fill the gap and provide a substantive base for creating effective prediction models.The study summarizes on a neural network method of financial credibility assessment, for which principal recommendations are developed with regard to the selection and possible variations of a neural network structure. Prediction models have been then synthesized for predicting inability to pay by Russian processing industry enterprises. The model has been tested against an analysis of financial position of Russian processing industry enterprises using their financial statements. Explanation is given for factors contributing to the enhanced accuracy of predictions made by the neural network model compared with the existing logistic regression-based models.Although no goal was set in the study to identify an optimal system of financial indicators for assessment of an enterprise’s financial credibility, the proposed approach can be applied in conjunction with any combination of financial indicators that ensures sufficient coverage of business aspects of an enterprise subjected to screening.
AU  - T Bogdanova
AU  - Timofey Shevgunov
AU  - Olga Uvarova
UR  - https://bijournal.hse.ru/en/2013--2(24)/90627823.html
PY  - 2013
SP  - 40-48
VL  -