| Title |
Power analysis of multivariate goodness of fit tests / |
| Authors |
Arnastauskaitė, Jurgita ; Ruzgas, Tomas ; Bražėnas, Mindaugas |
| DOI |
10.15388/DAMSS.12.2021 |
| ISBN |
9786090706732 |
| eISBN |
9786090706749 |
| Full Text |
|
| Is Part of |
DAMSS 2021: 12th conference on data analysis methods for software systems, Druskininkai, Lithuania, December 2–4, 2021.. Vilnius : Vilnius University press, 2021. p. 5.. ISBN 9786090706732. eISBN 9786090706749 |
| Abstract [eng] |
In modern data analytics, decisions making involves hypotheses testing. It is a common practice to check the assumption of data normality. Which dictates the choice of data analysis methods (parametric or non-parametric). The assumption of normality can be checked graphically, but a more consistent option is to test the goodness of fit hypothesis. Despite the fact that a lot of statistical test have been developed since the 20th century, analysis of multivariate data remains challenging. The purpose of this study is to perform a power analysis of multivariate goodness of fit hypothesis test for the assumption of normality for different data sets and to compare the results obtained with our proposed test. Thus, we proposed a new powerful multivariate test (MIDE), which is based on the mean absolute deviation of the empirical distribution density from the theoretical distribution density. In this test, the density estimate is derived by using a inversion formula. To show advantages of our test an exhaustive comparative study of multivariate tests was performed. For this purpose, a lot of multivariate data sets of non-normal distributions were generated. For the comparison, the power of well-known test and our test was evaluated empirically. Based on the obtained modelling results, it can be concluded that the MIDE test. |
| Published |
Vilnius : Vilnius University press, 2021 |
| Type |
Conference paper |
| Language |
English |
| Publication date |
2021 |
| CC license |
|
