Institutional Repository of Vilnius University
Finansinių rodiklių analizė ir jų taikymas bankroto rizikos lygio prognozavimui /
Abstract
dc:descriptionAssessing a company’s economic condition is important both for its managers – to enable business expansion – and for investors, who must decide which company to invest in based on financial indicators and projections. Constant monitoring of a company’s financial state also allows the timely detection of financial difficulties and the implementation of measures to prevent bankruptcy. The aim of this bachelor’s thesis is to analyse the financial indicators of Lithuanian companies, determine their interrelationships, and apply them in predicting the level of bankruptcy risk. The theoretical part reviews the literature and explores the methods commonly used in the analysis of financial indicators. In the practical part, an initial data analysis is conducted using factor analysis and principal component analysis (PCA) to identify the relationships among indicators. Altman Z and Springate S models are applied to determine the level of bankruptcy risk, and their results were used as input data for machine learning classifiers. Factor analysis yields three factors: the first factor, labelled "Finance", consists of revenue, total equity, accounts payable and other liabilities, and current liabilities. The second factor, "Profitability", includes EBIT and net profit. The third factor, "Capital Accumulation", is comprised of equity and current liabilities. Principal component analysis reveales three components explaining approximately 94% of the variance with a similar to factor analysis composition of components. Using the Altman Z model, it is determined that 125 companies had a low level of bankruptcy risk, 31 had a moderate level, and 92 had a high level. According to the Springate S model, 89 companies have a low level of bankruptcy risk, 16 have a moderate level, and 143 have a high level. Several machine learning classifiers are used to predict bankruptcy risk, including Random Forest, Support Vector Machine (SVM), k-Nearest Neighbours (k-NN), Logistic Regression, and XGBoost, where the latter is found the most accurate for both Atlman Z and Springate S models.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Dryžaitė, Ema,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD239028518&prefLang=en_US