DSpace - Tor Vergata >
Facoltà di Ingegneria >
Tesi di dottorato in ingegneria >

Please use this identifier to cite or link to this item: http://hdl.handle.net/2108/1048

Title: An Approach to improving parametric estimation models in the case of violation of assumptions based upon risk analysis
Authors: Cantone, Giovanni
Basili, Victor Robert
Sarcià, Salvatore Alessandro
Keywords: multi-layer feed-forward neural networks
non-linear regression
curvilinear component analysis
Bayesian learning
prediction intervals for neural networks
risk analysis and management
learning organizations
software cost prediction
TAME system
Bayesian discrimination function
estimation improvement paradigm
quality improvement paradigm
integrated software engineering environment
Issue Date: 27-Aug-2009
Abstract: In this work, we show the mathematical reasons why parametric models fall short of providing correct estimates and define an approach that overcomes the causes of these shortfalls. The approach aims at improving parametric estimation models when any regression model assumption is violated for the data being analyzed. Violations can be that, the errors are x-correlated, the model is not linear, the sample is heteroscedastic, or the error probability distribution is not Gaussian. If data violates the regression assumptions and we do not deal with the consequences of these violations, we cannot improve the model and estimates will be incorrect forever. The novelty of this work is that we define and use a feed-forward multi-layer neural network for discrimination problems to calculate prediction intervals (i.e. evaluate uncertainty), make estimates, and detect improvement needs. The primary difference from traditional methodologies is that the proposed approach can deal with scope error, m...
Description: 21. ciclo
URI: http://hdl.handle.net/2108/1048
Appears in Collections:Tesi di dottorato in ingegneria

Files in This Item:

File Description SizeFormat
SarciaPHDThesis.pdf1478KbAdobe PDFView/Open

Show full item record

All items in DSpace are protected by copyright, with all rights reserved.