Rui Américo Mathiasi Horta

Seleção de atributos na previsão de insolvência: aplicação e avaliação usando dados brasileiros recentes

ABSTRACT Bankruptcy prediction may have great utility to financial and nonfinancial institu-tions with regard to take in advance the best possible decisions regarding loans or investments. In specific literature, many bankruptcy prediction models have made use of data mining . The preprocessing step is important to select good qual-ity data for use in mining operations. Still, although the selection of attributes can be very beneficial to pre-select representative data to improve the forecast perfor-mance end, it is not known which method is the best selection. This work has as main objective to compare two approaches for evalua ting subsets of attributes: Filter and Wrapper. Despite being based on data mining techniques and widely used in the step of feature selection in bankruptcy prediction models, these tech-niques are rarely used to treat data from financial statements of Brazilian com-panies. Therefore the empirical basis of this study consists of a sample of Brazi-lian industrial and commercial enterprises, collecting data for the period 2004 to 2011. The results indicated that, in this sample, the filter approach was more effi-cient, providing better classification results both for logistic regression (91,80%) and for neural networks (93,98%). It was shown also the importance of making explicit the evaluation stage of the selection of attributes for achieving better re-sults in applications of data mining techniques to predict insolvency. A specific conclusion about the advantages of the filter approach shows that it may be pre-ferred to assess the attributes that will make predictive models.

Selección de atributos en previsión de insolvencia: aplicación y evaluación utilizando datos recientes de Brasil

ABSTRACT Bankruptcy prediction may have great utility to financial and nonfinancial institu-tions with regard to take in advance the best possible decisions regarding loans or investments. In specific literature, many bankruptcy prediction models have made use of data mining . The preprocessing step is important to select good qual-ity data for use in mining operations. Still, although the selection of attributes can be very beneficial to pre-select representative data to improve the forecast perfor-mance end, it is not known which method is the best selection. This work has as main objective to compare two approaches for evalua ting subsets of attributes: Filter and Wrapper. Despite being based on data mining techniques and widely used in the step of feature selection in bankruptcy prediction models, these tech-niques are rarely used to treat data from financial statements of Brazilian com-panies. Therefore the empirical basis of this study consists of a sample of Brazi-lian industrial and commercial enterprises, collecting data for the period 2004 to 2011. The results indicated that, in this sample, the filter approach was more effi-cient, providing better classification results both for logistic regression (91,80%) and for neural networks (93,98%). It was shown also the importance of making explicit the evaluation stage of the selection of attributes for achieving better re-sults in applications of data mining techniques to predict insolvency. A specific conclusion about the advantages of the filter approach shows that it may be pre-ferred to assess the attributes that will make predictive models.

Attribute selection in bankruptcy prediction: application and evaluation using recent brazilian data

ABSTRACT Bankruptcy prediction may have great utility to financial and nonfinancial institu-tions with regard to take in advance the best possible decisions regarding loans or investments. In specific literature, many bankruptcy prediction models have made use of data mining . The preprocessing step is important to select good qual-ity data for use in mining operations. Still, although the selection of attributes can be very beneficial to pre-select representative data to improve the forecast perfor-mance end, it is not known which method is the best selection. This work has as main objective to compare two approaches for evalua ting subsets of attributes: Filter and Wrapper. Despite being based on data mining techniques and widely used in the step of feature selection in bankruptcy prediction models, these tech-niques are rarely used to treat data from financial statements of Brazilian com-panies. Therefore the empirical basis of this study consists of a sample of Brazi-lian industrial and commercial enterprises, collecting data for the period 2004 to 2011. The results indicated that, in this sample, the filter approach was more effi-cient, providing better classification results both for logistic regression (91,80%) and for neural networks (93,98%). It was shown also the importance of making explicit the evaluation stage of the selection of attributes for achieving better re-sults in applications of data mining techniques to predict insolvency. A specific conclusion about the advantages of the filter approach shows that it may be pre-ferred to assess the attributes that will make predictive models.