Akademik Veri Yönetim Sistemi
Electronic Journal of Applied Statistical Analysis, cilt.7, sa.2, ss.254-278, 2014 (Scopus)
© Università del Salento.Functional t-tests involve testing for differences in functional means across two curve groups. If a significant overall difference in the mean curves is detected, one way to identify the location of these differences is pointwise testing. Ramsay and Silverman (2005) suggest using a pointwise test approach based on a permutation method and Cox and Lee (2008) based on Westfall-Young approach. Since both of them have their caveats, here it is suggested to use them simultaneously to improve the validation of results. It is concluded that they strongly support each other. The flexibility of functional data analysis is shown by using basis functions comparatively and the irregular behavior of the differences or mean functions across the years are presented via rainbow plots explicitly. Meteorological data like humidity is used for these aims. Functional t-tests are used in order to observe if humidity mean functions of coastal area and hinterland of Turkey are statistically ifferent. On the other hand, analysis for eleven years (2000-2010) is made in order to observe if these expectations changes in years. Suggested usage of the tests is proven to be practically useful.