DOI QR코드

DOI QR Code

Bivariate skewness, kurtosis and surface plot

이변량 왜도, 첨도 그리고 표면그림

  • Received : 2017.08.10
  • Accepted : 2017.09.18
  • Published : 2017.09.30

Abstract

In this study, we propose bivariate skewness and kurtosis statistics and suggest a surface plot that can visually implement bivariate data containing the correlation coefficient. The skewness statistic is expressed in the form of a paired real values because this represents the skewed directions and degrees of the bivariate random sample. The kurtosis has a positive value which can determine how thick the tail part of the data is compared to the bivariate normal distribution. Moreover, the surface plot implements bivariate data based on the quantile vectors. Skewness and kurtosis are obtained and surface plots are explored for various types of bivariate data. With these results, it has been found that the values of the skewness and kurtosis reflect the characteristics of the bivariate data implemented by the surface plots. Therefore, the skewness, kurtosis and surface plot proposed in this paper could be used as one of valuable descriptive statistical methods for analyzing bivariate distributions.

본 연구에서는 두 변수의 상관계수를 반영한 이변량 자료의 왜도와 첨도 통계량을 제안하고, 시각적으로 표현할 수 있는 표면그림을 개발한다. 이변량 왜도 통계량은 이변량 확률표본 자료의 치우침 방향과 정도를 표현하는 실수 한 쌍으로 정의한다. 첨도는 양의 값을 가지며 이변량 정규분포를 기준으로 꼬리 부분의 두터운 정도를 파악할 수 있다. 그리고 표면그림은 분위벡터를 바탕으로 평면에 구현한다. 다양한 형태의 이변량 자료를 생성하여 표면그림을 작성하고 왜도와 첨도를 계산하여 탐색해 본 결과, 왜도와 첨도 값들은 표면그림으로 구현한 이변량 자료의 특징을 잘 반영하는 것을 발견할였다. 그러므로 본 논문에서 제안한 왜도, 첨도 그리고 표면그림은 이변량 분포를 분석하는 기술통계학적 방법으로 활용할 수 있다.

Keywords

References

  1. Bickel, P. J. and Doksum, K. A. (2001). Mathematical statistics: Basic ideas and selected topics, 2nd Ed., Prentice-Hall, New Jersey.
  2. Hogg, R. V., Mckean, J. W. and Craig, A. T. (2013). Introduction to mathematical statistics, 7th Ed., Prentice-Hall, New Jersey.
  3. Hong, C. S. and Kim, H. J. (2016). Data analysis for regression model using SAS, Tamjin, Seoul.
  4. Hong, C. S., Han, S. J. and Lee, G. P. (2016). Vector at risk and alternative value at risk. The Korean Journal of Applied Statistics, 29, 689-697. https://doi.org/10.5351/KJAS.2016.29.4.689
  5. Hong, C. S. and Kim, J. Y. (2017). Multivariate cte for copula distributions. Journal of the Korean Data & Information Science Society, 28, 421-433. https://doi.org/10.7465/jkdi.2017.28.2.421
  6. Hong, C. S. and Lee, G. P. (2016). Properties of alternative var for multivariate normal distributions. Journal of the Korean Data & Information Science Society, 27, 1453-1463. https://doi.org/10.7465/jkdi.2016.27.6.1453
  7. Lehmann, E. L. and Casella, G. (1998). Theory of point estimation, 2nd Ed., Springer, New York.
  8. Lindgren, B. W. (1993). Statistical theory, 4th Ed., Chapman & Hall, London.
  9. Mardia, K. V. (1970). Measure of multivariate skewness and kurtosis with applications. Biometrica, 57, 519-530. https://doi.org/10.1093/biomet/57.3.519
  10. Mardia, K. V. (1974). Applications of some measures of multivariate skewness and kurtosis in testing normality and robustness studies. The Indian Journal of Statistics, Series B, 115-128.
  11. Mood, A. M., Graybill, F. A. and Boes, D. C. (1974). Introduction to the theory of statistics, 3rd ed., McGraw-Hill, New York.
  12. Rohatgi, V. K. (1976). An introduction to probability theory and mathematical statistics, John Wiley & Sons, New York.
  13. Seber, G. A. F. (1984). Multivariate observations, John Wiley & Sons, New York.