DISCONTINUANCE OF BRAZILIAN COMPANIES DO NOT CYCLIC CONSUMER SECTOR: A STUDY WITH FINANCIAL DATA USING DATA MINING TECHNIQUES

Abstract

The discontinuity of companies is an issue that increasingly is being studied in the field of accounting and finance due to the number of parts that are affected by the failure of a corporate entity be considerable. Banks, investors, governments, auditors, managers, suppliers, employees and many others have many interests in whether the reporting entity will continue to operate long enough to fulfill their commitments. In Brazil the studies in this area still have difficulties finding databases with reduced dimensions due to the quality of the data available. In this research the database is sourced from financial statements of Brazilian companies economic consumer staples listed on BOVESPA year 2000-2011. The research is classified according to purpose as a descriptive and applied. As for the means, it is an experimental research and documentary. It used data mining strategy that selects attributes and faces the issue of class imbalance, and this little problem studied in this subject because in normal economic environments the number of companies classified as solvents are much larger than those classified as insolvent. The research objective is to better characterize those companies belonging to the sector of consumer staples that have higher potential that they will become discontinuous attacking the problem of the imbalance between classes, in addition, by applying the selection of attributes. The results and validations performed the strategy presented demonstrate your success may well be considered competitive with other strategies presented in the specific literature.