Search results

1

Extreme gradient boosting method in the prediction of company bankruptcy

100%

Pawełek B.

Statistics in Transition new series

|

2019

|

vol. 20

|

issue 2

155-171

EN

Machine learning methods are increasingly being used to predict company bankruptcy. Comparative studies carried out on selected methods to determine their suitability for predicting company bankruptcy have demonstrated high levels of prediction accuracy for the extreme gradient boosting method in this area. This method is resistant to outliers and relieves the researcher from the burden of having to provide missing data. The aim of this study is to assess how the elimination of outliers from data sets affects the accuracy of the extreme gradient boosting method in predicting company bankruptcy. The added value of this study is demonstrated by the application of the extreme gradient boosting method in bankruptcy prediction based on data free from the outliers reported for companies which continue to operate as a going concern. The research was conducted using 64 financial ratios for the companies operating in the industrial processing sector in Poland. The research results indicate that it is possible to increase the detection rate for bankrupt companies by eliminating the outliers reported for companies which continue to operate as a going concern from data sets.

2

Outlier detection in the analysis of nested Gage R&R, random effect model

88%

Abduljaleel M., Midi H., Karimi M.

Statistics in Transition new series

|

2019

|

vol. 20

|

issue 3

31-56

EN

Measurement system analysis is a comprehensive valuation of a measurement process and characteristically includes a specially designed experiment that strives to isolate the components of variation in that measurement process. Gage repeatability and reproducibility is the adequate technique to evaluate variations within the measurement system. Repeatability refers to the measurement variation obtained when one person repeatedly measures the same item with the same Gage, while reproducibility refers to the variation due to different operators using the same Gage. The two factors factorial design, either crossed or nested factor, is usually used for a Gage R&R study. In this study, the focus is only on the nested factor, random effect model. Presently, the classical method (the method of analysing data without taking into consideration the existence of outliers) is used to analyse the nested Gage R&R data. However, this method is easily affected by outliers and, consequently, the measurement system’s capability is also affected. Therefore, the aims of this study are to develop an identification method to detect outliers and to formulate a robust method of measurement analysis of nested Gage R&R, random effect model. The proposed methods of outlier detection are based on a robust mm location and scale estimators of the residuals. The results of the simulation study and real numerical example show that the proposed outlier identification method and the robust estimation method are the most successful methods for the detection of outliers.

3

WYBRANE TESTY STATYSTYCZNE DLA WARTOŚCI NIETYPOWYCH I ICH ZASTOSOWANIE W ANALIZACH EKONOMETRYCZNYCH

63%

Pekasiewicz D.

Metody Ilościowe w Badaniach Ekonomicznych

|

2014

|

vol. 15

|

issue 4

111-120

PL

Wykrywanie obserwacji nietypowych w próbie losowej stanowi ważne zagadnienie w analizach statystycznych. Jednym ze sposobów badania próby od kątem istnienia wartości odstających jest stosowanie testów statystycznych opartych na statystykach ekstremalnych, do których należą: test Grubbsa i jego uogólnienie, test Dixona oraz testy oparte na asymptotycznych rozkładach minimum i maksimum z próby. Granicznymi rozkładami statystyk ekstremalnych są, w zależności od klasy rozkładu analizowanej zmiennej, rozkład Gumbela, Frecheta lub Weibulla. W artykule, oprócz rozważań teoretycznych, przedstawiono zastosowania wybranych testów do weryfikacji hipotez o wartościach nietypowych przy konstrukcji modeli ekonometrycznych.

EN

The problem of the existence of outliers in the sample is an important issue in statistical surveys. One of the methods of outliers detection is the application of statistical tests based on extreme statistics. Grubbs test and its generalization, Dixon test and tests based on asymptotic distributions of minimum and maximum (Gumbel, Frechet, Weibull distributions) belong to group of these tests. In the paper, besides the theoretical considerations the application of selected tests, used to verify the hypothesis of outliers in the construction of econometric models, is presented.

Refine search results

2 Statistics in Transition new series

1 Metody Ilościowe w Badaniach Ekonomicznych

1 Abduljaleel M.

1 Karimi M.

1 Midi H.

1 Pawełek B.

1 Pekasiewicz D.

2 2019

1 2014

Extreme gradient boosting method in the prediction of company bankruptcy

Outlier detection in the analysis of nested Gage R&R, random effect model

WYBRANE TESTY STATYSTYCZNE DLA WARTOŚCI NIETYPOWYCH I ICH ZASTOSOWANIE W ANALIZACH EKONOMETRYCZNYCH