Monte - Carlo evaluation of power transformations in regression model
2004
0 views
0 downloads
Advisor: Yrd. Doç. Dr. Mustafa Yavuz Ata
Abstract (EN)
In the classical regression model, it is assumed that the error terms are independently, identically and normally distributed. In this study, the cases where the assumptions of normality and homoskedasticity are not valid were modelled and the performances of the power transformations were compared by a Monte-Carlo virtual experiment. The results showed that Tukey's transformation and Manly's transformation performed better than others, respectively in case of the departure from normality and from homoskedasticity. Science Code : 212 Key Words Regression; power transformation; Monte-Carlo estimation Page Number : 64 Adviser : Ass. Prof. Dr. Mustafa Yavuz ATA
Author
Dr. Hüseyin Güler
How to Cite
Hüseyin Güler (Master Thesis). Monte - Carlo evaluation of power transformations in regression model, 2004, Gazi University.
Keywords
License
Tüm Hakları Saklıdır
This work is shared under the specified license terms.
More theses from Gazi University
- Occupational accident analysis and modelling in oil and gas drilling sector Turkey(2021)
- Experimental development of the interfacial bond-slip model between textile reinforced mortar strips and masonry walls(2025)
- Deveplopment of semiconductor humidity sensors(2021)
- The effect of computer-assisted and direct strategy teaching on reading comprehension(2021)
- Sharing of real life geometry samples via a social learning environment: A case study(2021)
- An investigation of the contribution of oud player to music education in terms of microphoning techniques, use of equalizer and sampler production(2021)