The aim of the first part of the project is to check the power of normality tests for data generated by simulation of three distributions: t-Student, uniform, and exponential. Five tests were used: Shapiro-Wilk, Jarque-Bera, Kolmogorov, Chi-quadrat and Anderson Darling. This part is based on writing a function calculating the power of the tests for distributions of a given length, and then visualizing the above.
In the second part of the project the ANOVA study was conducted to check whether the genre of the film, its length and the number of Oscars it won influence the rating of the film (one-way ANOVA), the combined influence of the film length and the Oscars on the rating of the film (two-way ANOVA) and the simultaneous influence of the Oscars on the rating of the film as well as on its length (MANOVA).
jakubig / anova_and_power_of_normality_tests Goto Github PK
View Code? Open in Web Editor NEWTesting the power of normality tests for data simulation and conducting both ANOVA (one-way and two-way) and MANOVA for my own movie dataset