Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Multivariate statistics”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 127 records · Page 7Linked to original sources
PubMedDate not supplied

Diagnostic effectiveness of serum bile acids in liver diseases as evaluated by multivariate statistical methods.

The aims of this study were to determine the diagnostic effectiveness of fasting and postprandial serum bile acid determinations in liver diseases, and to compare results with those of conventional liver function tests. In 322 patients with biopsy-proved liver disease and 93 healthy subjects, fasting and postprandial (2 hr) serum levels of cholic, chenodeoxycholic, and lithocholic acid conjugates and conventional liver function tests were evaluated. Data were subjected to variance and discriminant and factor analyses. Fasting serum bile acids were higher in patients when compared to controls and were significantly higher in severe than in mild liver diseases. Determination of cholic plus lithocholic acid provided the highest discrimination capacity. The percent of correct allocation was 75.4% for conventional liver function tests, 70.1% for fasting serum bile acids and increased to 79.6% when liver function tests plus serum bile acids were considered. Postprandial percentages were always lower than fasting. Factor analysis identified two factors possibly related to cytolysis and protein synthesis. The serum bile acid concentrations highly correlated with both factors. We conclude that serum bile acid determinations increase the diagnostic and discriminant capacities of liver function tests and are more sensitive and discriminant when obtained in fasting than postprandially.

Adult↗
PubMed1986

[The good use of multivariate statistics].

Within a multidisciplinary team, it is important to establish a common language for the statistician and the electroencephalographer. Using an example, we propose a brief introduction to correspondence analysis. Within a set of 8 variables, we show how to isolate a subset of variables which are characteristic either for partial epilepsy or generalized epilepsy.

Adult↗
PubMedDate not supplied

Multivariate statistical methods in toxicology. III. Specifying joint toxic interaction using multiple regression analysis.

Multiple regression is widely employed to study the contribution of components to the toxicologic effect of a mixture. Here, use is made of the fact that data obtained from standard curves of substances and from their mixtures are separable in regression analysis. Thus, under an assumption of additivity of responses, regression coefficients obtained for components in mixtures alone should be the same as for the individual substances. A t-test is developed such that nonsignificant t values support additivity, negative significant values support antagonism, and positive significant values support synergism. The results are applied to data on the mutagenicity of binary mixtures of azaserine, 4-nitroquinoline N-oxide, and 9-aminoacridine in TA 100 in the Ames assay.

4-Nitroquinoline-1-oxide↗