. "0"^^ . "GA13-01930S" . . . "2016-12-31+01:00"^^ . "1"^^ . . " robust tests of orthogonality" . "0"^^ . "4"^^ . "Robustifikovan\u00E9 (vzhledem ke kontaminaci dat) verze ekonometrick\u00FDch metod by m\u011Bly realisticky p\u0159edpokl\u00E1dat charakter ekonomick\u00FDch dat, nap\u0159. korelace mezi regresory a disturbancemi t\u00E9m\u011B\u0159 automaticky implikuje heteroskedasticitu a kolinearita zase m\u016F\u017Ee maskovat kontaminaci. Robustifikovan\u00E9 metody by tedy m\u011Bly fungovat i p\u0159i ``kombinovan\u00FDch naru\u0161en\u00EDch\u2018\u2018 klasick\u00FDch p\u0159edpoklad\u016F. N\u011Bkter\u00E9 takov\u00E9 nestandardn\u00ED situace nebyly dosud dostate\u010Dn\u011B prostudov\u00E1ny s c\u00EDlem navr\u017Een\u00ED robustn\u00EDch n\u00E1stroj\u016F, kter\u00E9 si s nimi porad\u00ED, jako je model s bin\u00E1rn\u00ED \u010Di limitovanou z\u00E1visl\u00E9 prom\u011Bnnou, modely uva\u017Euj\u00EDc\u00ED censorovan\u00E1 \u010Di chyb\u011Bj\u00EDc\u00ED data. Zat\u00EDm byly nalezeny jen robustifikovan\u00E9 verze n\u011Bkter\u00FDch diagnostick\u00FDch n\u00E1stroj\u016F, zejm\u00E9na pro nejmen\u0161\u00ED v\u00E1\u017Een\u00E9 \u010Dtverce a instrument\u00E1ln\u00ED v\u00E1\u017Een\u00E9 prom\u011Bnn\u00E9. Diagnostick\u00E9 n\u00E1stroje mus\u00ED toti\u017E b\u00FDt robustifikov\u00E1ny individu\u00E1ln\u011B pro ka\u017Edou robustifikovanou metodu (p\u016Fvodn\u011B byly tyto konstruov\u00E1ny pro N\u010C). Algoritmizace, implementace a odzkou\u0161en\u00ED jsou sice \u00FAnavn\u00E9 pr\u00E1ce, ale bez nich ztr\u00E1c\u00ED na\u0161e v\u00FDsledky atraktivitu a spolehlivost v aplikac\u00EDch. To implikuje, co bychom m\u011Bli studovat." . "4"^^ . . . " efficiency of robust analysis" . . "2015-04-23+02:00"^^ . " models with effects and qualitative response" . . . " significance of regressors" . " algorithms" . "Robust methods for nonstandard situations, their diagnostics and implementations"@en . . " homoscedasticity" . "Regression model, robust estimators , diagnostics, robust tests of orthogonality, homoscedasticity, collinearity, models with effects and qualitative response, efficiency of robust analysis, significance of regressors, algorithms, their implementations \u2026"@en . "Robustified (with respect to contamination) versions of econometric methods have to consider realistically the character of economic data, e. g. correletion of regressors and disturbances nearly automatiaclly implies heteroscedastcity or collinearity can mask contamination. So, the robustified versions are to work under ``combined\u2019\u2019 distortion of classical assumptions \u2013i. e. under nonstandard situations. Some nonstandard situations were not yet exhaustively studied with a goal of robustified tool for coping with them, as binary or limited response, even worse situation is with the censored and missing data. Only some basic diagnostic tools were considered for robustified regression model identification methods, namely for least weighted squares and instrumental weighted variables. Diagnostic tools have to be tailored individually for every robustified method (as they were tailored for the OLS in the past). Algorithmization, implementation and verification are tiresome work without which the theoretical findings lose on reliability and attractivity. It implies what is to be done."@en . . . . " robust estimators " . "http://www.isvav.cz/projectDetail.do?rowId=GA13-01930S"^^ . "2014-03-31+02:00"^^ . . . " diagnostics" . . . "2013-02-01+01:00"^^ . "Regression model" . "Robustn\u00ED procedury pro nestandardn\u00ED situace, jejich diagnostika a implementace" . " collinearity" .