| Abstract [eng] |
Corporate bankruptcies in Lithuania was a common phenomenon in the market economy, causing many negative consequences not only for the company, but also for the state and society. In order to seek to ensure the success of businesses, their managers must be able to assess their business risks and make effective decisions in order to avoid bankruptcies threats. Financial data research is the easiest way to assess the company's business continuity. Company bankruptcy rudiments and its threat can be observed in the context of the financial statements of the changes in calculating and comparing the relative financial indicators, analyzing their dynamics, using the bankruptcy prediction models. Classic bankruptcy prediction models are commonly used in the examination Lithuanian companies’ bankruptcy, but many scientists recognize that they deliver results with a large margin of error, which leads to the conclusion that these models obsolete. Therefore, to better understand whether the is company at risk of bankruptcy, and what it leads to is worth a study use the modern bankruptcy prediction models, which are not yet widely adapted in Lithuania. Purpose - to make a research of bancruptcy prediction methods' application for Lithuanian companies. The object - bankruptcy prediction models. Objectives: 1. Discuss the classical, modern and updated classic bankruptcy prediction models in Lithuanian companies’ bankruptcy factors to predict problems. 2. Analyze what bankruptcy prediction methods, scientists propose to use to determine the factors influencing bankruptcy Lithuania. 3. The bankruptcy prediction methods and methodology. 4. Perform bankruptcy prediction models in Lithuanian enterprises investigation. Classic bankruptcy prediction models introduced since a long time ago. Widely used discriminant analyzes bankruptcy prediction models predict the likelihood of bankruptcy the same or similar places and can be applied to Lithuanian companies. However, the applicability of these techniques is very critical of many Lithuanian scientists as classical models can not fully predict whether the company in danger of bankruptcy, they can only complement the set of methods, which should be built precisely Lithuanian and Baltic market. Therefore, many of the latest researchers analyzed bankruptcy prediction methods Lithuanian market, said that the classic models can only be a complementary part of the analysis, so more and more scientists advise to try modernity bankruptcy prediction methods, and follow the company's indicators, which may indicate the company financiers bankruptcy risk. The survery results showed that the most accurate and most reliable classic models Lithuanian corporate bankruptcy probability to predict the linear discriminant analysis group are patterns and predicts the true financial conditions of companies. Logistic regression models results are controversial. But scientists confirmed that more reliable to choose modernity bankruptcy models to predict bankrupcy to Lithuanian companies. One of the modern bankruptcy prediction model - neural networks in the study helps to clarify the application of the challenges bankruptcy set. Bankruptcy forecasting models during the investigation was carried out screening of companies, it has also been studies determining bankruptcy adapting classical neural network and integrating neural kinds of network interfaces bankruptcy prediction models. And present and compare the results obtained using different bankruptcy prediction models for investigating Lithuanian market corporate bankruptcy. The survey results showed that neural networks can to precisely predict Lithuanian companies bankruptcy. Also we found that classical models is the greatest accuracy of Altman Chesser models. The survery have shown that precisely bankruptcy predicted combinated neural kinds of network interfaces model. In a few percentage points select a lower probabilities predicted the relative performance of the neural network. So it can be concluded that only classical bankruptcy prediction model is not appropriate to use in predicting bankruptcy. The correct choice of the relative indicators and combining the results with these indicators Altman and Chesser model and have been trained neural network, it is possible to get more than ninety percent probability that the network will predict reliably companies operating condition of the coming year. |
