Community Research and Development Information Service - CORDIS


This work deals with some aspects of the statistical analysis of complex technological systems. Its aim is the formulation of a final judgement on the expected stochastic behaviour of a given system, when information is mainly of a possibly subjective nature and collected through different channels: elicitation of experts, simulation and experimented data. The model assumed for the system's responses is a regression equation relating the (nx1) vector of observations to a (nxk) non-stochastic design matrix, a (kx1) parameter vector and the (nx1) error vector, assumed to be normally distributed. The problem discussed is how to fit a pooling of information, having different credibility degrees, on a unique interpretation model and how to assess the importance of each variable.

Additional information

Authors: CALVI PARISETTI C, University of Parma, Department of Mathematics (IT);OLIVI L, JRC Ispra (IT);PESARIN F, University of Padua, Department of Statistics (IT)
Bibliographic Reference: Paper presented: Probabilistic Safety Assessment and Management (PSAM), Beverley Hills, California (US), Feb. 4-7, 1991
Availability: Available from (1) as Paper EN 35603 ORA
Record Number: 199011727 / Last updated on: 1994-12-02
Original language: en
Available languages: en