Within the Research and Development activities, complex statistical problems of hypothesis testing can commonly arise. The complexity of the problem is mainly referred to the multivariate nature of the study and possibly to the presence of mixed performance variables (ordinal categorical, binary or continuous) and sometimes to missing values as well. In this contribution we consider permutation methods for multivariate testing on mixed variables within the framework of multivariate randomised complete block design. The novel approach we propose has been studied and validated via Monte Carlo simulation study. Finally we propose an application to real data, where several panellists from an R&D division of an home-care company are enrolled to studying several possible new fragrances of a given detergent to be compared with the own presently marketed product.
Future Trends on Global Performance Indicators in Industrial Research
CORAIN, LIVIO;SALMASO, LUIGI
2009
Abstract
Within the Research and Development activities, complex statistical problems of hypothesis testing can commonly arise. The complexity of the problem is mainly referred to the multivariate nature of the study and possibly to the presence of mixed performance variables (ordinal categorical, binary or continuous) and sometimes to missing values as well. In this contribution we consider permutation methods for multivariate testing on mixed variables within the framework of multivariate randomised complete block design. The novel approach we propose has been studied and validated via Monte Carlo simulation study. Finally we propose an application to real data, where several panellists from an R&D division of an home-care company are enrolled to studying several possible new fragrances of a given detergent to be compared with the own presently marketed product.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
