Využití analýzy přežití k predikci bankrotu

Abstract

The survival analysis is complex of statistical methods, which tries to determine the time between two events. The necessary data for this analysis are initial and terminate event and independent covariates, that have impact to if and when the terminate event occurs. In this thesis is described application in finance and it is done by examination of time to bankruptcy of given subject. The aim of the thesis is evaluation of survival time of the Czech companies from their foundation day to their bankruptcy. First of all are compared survival functions and hazard functions for particular sectors of Czech economic. Analyzed are the companies from administration, information technology, culture, building industry, accommodation, health care, agriculture and water and waste treatment. The source for the analysis are firms, in that the bankruptcy did not occur and also the firms, in that the bankruptcy occurred between years 2005 and 2015. In the followed part of the thesis is built Cox semiparametric model and parametric model in condition that survival time corresponds to Weibull probability distribution. The goal of these models estimation is to identify the financial ratios, which influent time to company bankruptcy and to quantify this impact.

Description

Import 02/11/2016

Subject(s)

Survival analysis, Bankruptcy, Censoring, Cox model, Survival function, Hazard function, Kaplan-Meier estimator, Logrank test, Nelson-Aalen estimator, Nonparametric methods, Parametric model, Likelihood function, Weibull distrbution, Wilcoxon test

Citation