Location: Cascina, Italy
Member since: May 2012
Duferco Energia S.p.A.
Development of automated and efficient VBA sub-routines for the delivery of data, establishing connections between Excel, Word and Outlook to produce reporting and analysis to support decision making.
- Gas consumptions forecasting;<br />- Computation of necessary gas capabilities to satisfy customers' demand in each area;<br />- Creation of VBA routines for data analysis;<br />- Database Access update through query;<br />- Creation of reports and statistical analyses on Excel (via pivot tables, array formulae, macro);<br />- Support for corporate processes mapping and definition of the informative data flow.
Prysmian Cavi e Sistemi Italia s.r.l.
- Development of multivariate statistical models (logistic regression, principal components analysis, Monte Carlo and cross-validation simulations, ROC curve analysis) to correlate the percentage of scraps with the parameters of the raw materials, using the R software;<br />- Production scheduling: creation of Excel reports about produced items, buffers, stocks.
National Institute of Geophysics and Volcanology
- Development of statistical models about expert elicitation techniques;<br />- Development of new algorithms using the software R;<br />- Multivariate statistical analyses (cluster analysis, principal components analysis, regression,) about volcanic eruptions data sets;<br />- Creation of probabilistic maps about volcanic risk management.
Università di Pisa
Comparison of a new expert elicitation model with the Classical Model, equal weights and single experts, using a cross-validation technique
Reliability Engineering and System Safety
The problem of ranking and weighting experts' performances when quantitative judgments are being<br />elicited for decision support is considered. A new scoring model, the Expected Relative Frequency<br />model, is presented, based on the closeness between central values provided by the expert and known<br />values used for calibration. Using responses from experts in five different elicitation datasets, a crossvalidation<br />technique is used to compare this new approach with the Cooke Classical Model, the Equal<br />W