2 resultados para Statistical Power

em DI-fusion - The institutional repository of Université Libre de Bruxelles


Relevância:

60.00% 60.00%

Publicador:

Resumo:

Treatment regimens for solid tumours have been extensively investigated for their physical toxic effects, but far less is known about the potential impairment of cognitive function by anticancer treatment regimens. Here, we review published studies that examined cognitive function in adult patients receiving systemic therapy for solid tumours. Our review suggests that patients can experience cognitive changes related to their treatment. However, several studies had methodological limitations, such as use of a limited sample size, lack of baseline assessment, and lack of control for potential confounding factors. Better designed clinical trials are required so that the difficulties patients face in terms of reduced cognitive function as a result of anticancer treatment can be fully elucidated. These trials should have sufficient statistical power and, importantly, should also be prospective.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

Static state estimators currently in use in power systems are prone to masking by multiple bad data. This is mainly because the power system regression model contains many leverage points; typically they have a cluster pattern. As reported recently in the statistical literature, only high breakdown point estimators are robust enough to cope with gross errors corrupting such a model. This paper deals with one such estimator, the least median of squares estimator, developed by Rousseeuw in 1984. The robustness of this method is assessed while applying it to power systems. Resampling methods are developed, and simulation results for IEEE test systems discussed. © 1991 IEEE.