"Metody statistick\u00E9 inference zalo\u017Een\u00E9 na matici p\u00E1rov\u00FDch vzd\u00E1lenost\u00ED"@cs . . "661"^^ . . "2013-12-31+01:00"^^ . . "661"^^ . . "0"^^ . "5"^^ . "0"^^ . "5"^^ . . "2011-01-01+01:00"^^ . "distance matrix; metrics; testing of statistical hypothesis"@en . . "Many biological and ecological data sets have a very large number of variables, which are highly skewed, and zero-inflated. Thus data analysts often do not rely on parametric assumptions, but calculate a distance matrix, which is the basis for further statistical inference. The most common situation is an ANOVA design and our interest is in testing for an effect of a treatment on outcome. The standard method based on a distance matrixis the ANOSIM (analysis of similarities) procedure which can be viewed as a multivariate ANOVA method based on a distance matrix. Recently, a method focused on testing the homogeneity of dispersion has been also suggested. The main objective of the project is (1) to investigate and compare various version of ANOSIM test that have appeared in the literature, (2) to explore the properties of the suggested test for dispersions and develop alternative methods, (3) to utilize the knowledge gained in (1) and (2) in more complex designs than ANOVA."@en . "1"^^ . . "Methods of statistical inference based on a distance matrix"@en . . "Data poch\u00E1zej\u00EDc\u00ED z biologick\u00FDch \u010Di ekologick\u00FDch aplikac\u00ED maj\u00ED \u010Dasto velmi mnoho prom\u011Bnn\u00FDch, kter\u00E9 maj\u00ED se\u0161ikmen\u00E9 rozd\u011Blen\u00ED a obsahuj\u00ED mnoho nul. Proto se p\u0159i anal\u00FDze takov\u00FDch dat nespol\u00E9h\u00E1me na p\u0159edpoklady parametrick\u00FDch model\u016F, ale spo\u010Dte se matice vzd\u00E1lenost\u00ED, kter\u00E1 je v\u00FDchoz\u00EDm bodem n\u00E1sledn\u00E9 statistick\u00E9 inference. Nej\u010Dast\u011Bj\u0161\u00ED situac\u00ED je anal\u00FDza jednoduch\u00E9 t\u0159\u00EDd\u011Bn\u00ED, kdy n\u00E1s zaj\u00EDm\u00E1, zda o\u0161et\u0159en\u00ED/faktor m\u00E1 vliv na sledovanou veli\u010Dinu. Standardn\u00ED metodou zalo\u017Eenou na matici vzd\u00E1lenost\u00ED je v tomto p\u0159\u00EDpad\u011B ANOSIM (anal\u00FDza podobnost\u00ED), na kterou se m\u016F\u017Eeme d\u00EDvat jako na druh v\u00EDcerozm\u011Brn\u00E9 anal\u00FDzy jednoduch\u00E9ho t\u0159\u00EDd\u011Bn\u00ED. Ned\u00E1vno byla tak\u00E9 navr\u017Eena metoda, kter\u00E1 se zam\u011B\u0159uje na testov\u00E1n\u00ED homogenity disperze. Hlavn\u00EDm c\u00EDlem projektu je (1) prozkoumat a porovnat r\u016Fzn\u00E9 verze ANOSIM testu, kter\u00E9 se objevily v literatu\u0159e, (2) prozkoumat vlastnosti testu pro disperzi a navrhnout alternativn\u00ED metody, (3) vyu\u017E\u00EDt znalosti z bod\u016F (1) a (2) pro slo\u017Eit\u011Bj\u0161\u00ED pl\u00E1ny ne\u017E anal\u00FDza jednoduch\u00E9ho t\u0159\u00EDd\u011Bn\u00ED."@cs . . . .