. "CZ - \u010Cesk\u00E1 republika" . . "RIV/00216275:25310/07:00006248" . . "Computer-Assisted Statistical Data Analysis. 6. Multivariate Classification of Various Sources of Drinkable Water using Principal Component Analysis and Cluster Analysis"@en . "Vodn\u00ED hospod\u00E1\u0159stv\u00ED" . . . . . "V\u00EDcerozm\u011Brn\u00E1 statistick\u00E1 anal\u00FDza je zalo\u017Eena na latentn\u00EDch prom\u011Bnn\u00FDch, kter\u00E9 jsou line\u00E1rn\u00ED kombinac\u00ED p\u016Fvodn\u00EDch prom\u011Bnn\u00FDch, y = w1 x1+...+wm xm . Zdrojov\u00E1 matice dat obsahuje prom\u011Bnn\u00E9 v m sloupc\u00EDch a objekty v n \u0159\u00E1dc\u00EDch. Data jsou p\u0159ed zpracov\u00E1n\u00EDm \u0161k\u00E1lov\u00E1na. C\u00EDlem je nal\u00E9zt shluk jako mno\u017Einu podobn\u00FDch objekt\u016F s podobn\u00FDmi prom\u011Bnn\u00FDmi. Podobnost objekt\u016F posuzujeme na z\u00E1klad\u011B vzd\u00E1lenosti (m\u00EDry) objekt\u016F v m-rozm\u011Brn\u00E9m prostoru: \u010D\u00EDm je vzd\u00E1lenost shluk\u016F \u010Di objekt\u016F v\u011Bt\u0161\u00ED, t\u00EDm men\u0161\u00ED je jejich podobnost. Strukturu a vazby mezi prom\u011Bnn\u00FDmi vystihuj\u00ED metody sn\u00ED\u017Een\u00ED dimensionality, metoda hlavn\u00EDch komponent (PCA). D\u016Fle\u017Eitou pom\u016Fckou je rozptylov\u00FD diagram, kter\u00FD zobrazuje objekty, rozpt\u00FDlen\u00E9 v rovin\u011B prvn\u00EDch dvou hlavn\u00EDch komponent. Graf komponentn\u00EDch vah porovn\u00E1v\u00E1 vzd\u00E1lenosti mezi prom\u011Bnn\u00FDmi xi a xj, kde kr\u00E1tk\u00E1 vzd\u00E1lenost zna\u010D\u00ED silnou korelaci. Dvojn\u00FD graf pak kombinuje oba p\u0159edchoz\u00ED grafy. Objekty lze seskupovat do shluk\u016F hierarchicky dle p\u0159edem zvolen\u00E9ho zp\u016Fsobu metriky (pr\u016Fm\u011Brov\u011B, centroidn\u011B, nejbli\u017E\u0161\u00EDm sousedem" . "289-296" . "Statistick\u00E9 zpracov\u00E1n\u00ED vodohospod\u00E1\u0159sk\u00FDch dat 6. V\u00EDcerozm\u011Brn\u00E1 klasifikace zdroj\u016F pitn\u00E9 vody metodou hlavn\u00EDch komponent PCA a shluk\u016F CLU" . . . "57" . "V\u00EDcerozm\u011Brn\u00E1 statistick\u00E1 anal\u00FDza je zalo\u017Eena na latentn\u00EDch prom\u011Bnn\u00FDch, kter\u00E9 jsou line\u00E1rn\u00ED kombinac\u00ED p\u016Fvodn\u00EDch prom\u011Bnn\u00FDch, y = w1 x1+...+wm xm . Zdrojov\u00E1 matice dat obsahuje prom\u011Bnn\u00E9 v m sloupc\u00EDch a objekty v n \u0159\u00E1dc\u00EDch. Data jsou p\u0159ed zpracov\u00E1n\u00EDm \u0161k\u00E1lov\u00E1na. C\u00EDlem je nal\u00E9zt shluk jako mno\u017Einu podobn\u00FDch objekt\u016F s podobn\u00FDmi prom\u011Bnn\u00FDmi. Podobnost objekt\u016F posuzujeme na z\u00E1klad\u011B vzd\u00E1lenosti (m\u00EDry) objekt\u016F v m-rozm\u011Brn\u00E9m prostoru: \u010D\u00EDm je vzd\u00E1lenost shluk\u016F \u010Di objekt\u016F v\u011Bt\u0161\u00ED, t\u00EDm men\u0161\u00ED je jejich podobnost. Strukturu a vazby mezi prom\u011Bnn\u00FDmi vystihuj\u00ED metody sn\u00ED\u017Een\u00ED dimensionality, metoda hlavn\u00EDch komponent (PCA). D\u016Fle\u017Eitou pom\u016Fckou je rozptylov\u00FD diagram, kter\u00FD zobrazuje objekty, rozpt\u00FDlen\u00E9 v rovin\u011B prvn\u00EDch dvou hlavn\u00EDch komponent. Graf komponentn\u00EDch vah porovn\u00E1v\u00E1 vzd\u00E1lenosti mezi prom\u011Bnn\u00FDmi xi a xj, kde kr\u00E1tk\u00E1 vzd\u00E1lenost zna\u010D\u00ED silnou korelaci. Dvojn\u00FD graf pak kombinuje oba p\u0159edchoz\u00ED grafy. Objekty lze seskupovat do shluk\u016F hierarchicky dle p\u0159edem zvolen\u00E9ho zp\u016Fsobu metriky (pr\u016Fm\u011Brov\u011B, centroidn\u011B, nejbli\u017E\u0161\u00EDm sousedem"@cs . "Statistick\u00E9 zpracov\u00E1n\u00ED vodohospod\u00E1\u0159sk\u00FDch dat 6. V\u00EDcerozm\u011Brn\u00E1 klasifikace zdroj\u016F pitn\u00E9 vody metodou hlavn\u00EDch komponent PCA a shluk\u016F CLU" . "8"^^ . "[066F3A8D27E5]" . . "1211-0760" . "25310" . . "452374" . . . "PCA; Principal Components Analysis; Cluster Analysis; Dendrogram; Drinkable Water; Water analysis; Potable water; Scatterplot; Scree Plot; Components Weight Plot; Correlation matrix."@en . "Meloun, Milan" . "RIV/00216275:25310/07:00006248!RIV08-MSM-25310___" . . . "Statistick\u00E9 zpracov\u00E1n\u00ED vodohospod\u00E1\u0159sk\u00FDch dat 6. V\u00EDcerozm\u011Brn\u00E1 klasifikace zdroj\u016F pitn\u00E9 vody metodou hlavn\u00EDch komponent PCA a shluk\u016F CLU"@cs . . . "Statistick\u00E9 zpracov\u00E1n\u00ED vodohospod\u00E1\u0159sk\u00FDch dat 6. V\u00EDcerozm\u011Brn\u00E1 klasifikace zdroj\u016F pitn\u00E9 vody metodou hlavn\u00EDch komponent PCA a shluk\u016F CLU"@cs . "1"^^ . . . "Multivariate statistical analysis is based on the latent variables which are formed as the linear combination of original variables y = w1 x1+...+wm xm . Data matrix contains objects in n rows and m columns. Before data treatment the data are scaled. Similarity of objects and variables is considered on base on Mahalonobis distance or Euclidean distance in the mdimensional space. The principal components analysis reduces dimensionality and presents objects in two or three dimensions. The plot of components weight shows hidden structure among variables while the scatterplot shows the hidden structure of objects. The cluster analysis leads to clusters which may be plotted in dendrogram. There are two dendrograms available, the dendrogram of variables and the dendrogram of objects. Both statistical techniques are demonstrated on the analysis and classification of various sources of a drinkable water."@en . "Computer-Assisted Statistical Data Analysis. 6. Multivariate Classification of Various Sources of Drinkable Water using Principal Component Analysis and Cluster Analysis"@en . "1"^^ . "8" . . "Z(MSM0021627502)" . . . .