ResearchBib Share Your Research, Maximize Your Social Impacts
Sign for Notice Everyday Sign up >> Login

Using Kohonen's Neural Networks to Identify the Bankruptcy of Enterprises: Case Study Based on Construction Companies in South Bohemian Region

Proceeding: 5th International Conference on Innovation Management, Entrepreneurship and Sustainability (IMES)

Publication Date:

Authors : ;

Page : 985-995

Keywords : construction industry; Kohonen's networks; bankruptcy;

Source : Downloadexternal Find it from : Google Scholarexternal


Kohenen's neural networks are used for the creation of cluster analyses. Through these networks, we search for common features within a data set and then we interpret these, ideally with validity for the entire data set. The aim of the article is to conduct a cluster analysis on a data set concerning construction enterprises in the Southern Bohemian region. Accordingly, we will identify the basic variables common to companies that enter liquidation.Design/methodology/approach: The data set includes financial statements and non-financial information (e.g. number of employees, business location) of the construction companies, all of which had been in operation from 2006 to 2015 in the Southern Bohemian Region. The data was organized into a table where one row contained information about one company in a particular year. Subsequently, Kohonen's neural networks were generated and company clusters identified. We then focused attention on the clusters of companies that were in a state of liquidation and results of comparison then showed which sorting characteristics appeared repeatedly in the clusters of the companies.Findings: The Kohonen neural network was determined. After its application, we receive Kohenen's map. As a result, we have clusters of enterprises with the same characteristics. Subsequently, comparing the clusters in which there are companies in liquidation, we obtain the characteristics typical for companies experiencing difficulties.Research/practical implications: The practical outcome are the identified variables crucial for the survival of construction companies in the Southern Bohemian Region. The companies will thus be able to focus on the key factors of their success, or the indicators of their eventual failure. Their management will therefore be more efficient and accurate.Originality/value: The additional value is the use of neural networks and cluster analysis tools for predicting the possible bankruptcy of companies. The results in the form of identified key factors leading to the failure of construction companies in the Southern Bohemian Region also have their value.

Last modified: 2017-09-02 23:13:50