Застосування регресійних моделей для аналізу і прогнозування показників якості фінансової діяльності підприємства

• Дата: 2020
• Автори: Kuznietsova, Nataliia V., Chernysh, Zlata S.
• Формат: Стаття
• Мова: Ukrainian
• Опубліковано: The National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" 2020
• Теми: regression models; seasonal autoregression model with integrated moving average; linear multiple regression; data processing; heteroskedastic models; регресійні моделі; сезонна модель авторегресії з інтегрованим ковзним середнім; лінійна множинна регресія; попереднє оброблення даних; гетероскедастичні моделі; регрессионные модели; сезонная модель авторегрессии с интегрированным скользящим средним; линейная множественная регрессия; предварительная обработка данных; гетероскедастические модели
• Онлайн доступ: http://journal.iasa.kpi.ua/article/view/216208
• Назва журналу: System research and information technologies

The company's success forecasting problem based on its financial indicators by regression models was studied in this research. Models based on linear multiple regression, autoregression with moving average, autoregression with integrated moving average, and seasonal model of autoregression with integrated moving average were built to predict the absolute value of financial indicators. An experimental study was performed on real data, and forecasting was made based on regression models. The models based on the method of group method of data handling and autoregressive neural network were developed. Heteroskedastic models with variable volatility such as ARCH and GARCH type were used to predict the volatility of the financial series. Preliminary data processing using the Holt-Winters method and the Kalman filter were applied to improve the model's quality and forecasting accuracy significantly. Authors suggested and developed a combination of seasonal autoregression with integrated moving average and heteroskedastic models that allowed them to consider the seasonal effects and trends inherent in the financial series and obtain high forecasts for financial indicators.