A FUZZY DECISION SUPPORT SYSTEM FOR THE ENVIRONMENTAL RISK ASSESSMENT OF GENETICALLY MODIFIED ORGANISMS F. Camastra, A. Ciaramella, V. Giovannelli, M. Lener, V. Rastelli, A. Staiano, G. Staiano, A. Starace WIRN 2013 - XXIII Italian Workshop on Neural Networks Vietri sul Mare, Salerno, Italy May 23 rd 2013
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
A FUZZY DECISION SUPPORT SYSTEM FOR THE
ENVIRONMENTAL RISK ASSESSMENT OF GENETICALLY
MODIFIED ORGANISMS F. Camastra, A. Ciaramella, V. Giovannelli, M. Lener, V. Rastelli, A. Staiano, G. Staiano, A. Starace
WIRN 2013 - XXIII Italian Workshop on Neural Networks
Vietri sul Mare, Salerno, ItalyMay 23rd 2013
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, Italy
Contents
Introduction Environmental Risk Assessment (ERA) Fuzzy Decision Support System System Validation Conclusions and Future Works
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Introduction
Genetically Modified Organism (GMO) organism altered using genetic engineering
techniques
ADVANTAGES GM Crop plant
resistant to herbicide, pests, diseases or environmental conditions
with improved nutritional or pharmaceutical properties
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
BUT …
Risks that might impact on: Consumers
Man, Animals (e.g., butterflies) Natural environment
Natural habitats, Soil
European directives assess and manage GMO risks The notifier, i.e., the person who
requests the GMO release, must perform an ERA
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Environmental Risk Assessment (ERA)1. Preliminary identification of risks
available scientific database and literature
2. Effects on non-target species BT toxin and butterflies
3. Effects on natural environment Compatible wild plants
4. Management of the risk mitigation measures
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
The Conceptual Model
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Overview of ERA process
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
System architecture
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
The Electronic Questionnaire
As a web application Different kinds of questions:
Textual Numerical Date Linguistic Multiple choice
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Textual question
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Numerical question
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Date question
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Linguistic question
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Multiple choice question
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
The Fuzzy Decision Support System (FDSS)
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Fuzzifier
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Fuzzy rule base
IF vegetative cycle duration IS HighAND cultural cycle duration IS LowTHEN phenological risk IS Low
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Inference engine
IF vegetative cycle duration IS HighAND cultural cycle duration IS LowTHEN phenological risk IS Low
IF vegetative cycle duration IS HighAND cultural cycle duration IS HighTHEN phenological risk IS High
…
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Defuzzifier
IF vegetative cycle duration IS HighAND cultural cycle duration IS LowTHEN phenological risk IS Low
IF vegetative cycle duration IS HighAND cultural cycle duration IS HighTHEN phenological risk IS High
…
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Report
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Open source library http://jfuzzylogic.sourceforge.net/html/index.html
Fuzzy Control Language (FCL) specification
Aggregation, Activation and Accumulation methods
Supports continue, discrete or custom membership functions
Flexible and extensible FDSS
jFuzzyLogic
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
System validation
The Fuzzy Decision Support System has been tested by producing about 150 ERAs by submitting the produced inferences to a
pool of ISPRA experts not involved in the rule definition
to assess the consistency and completeness of the system
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Scenario 1
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Scenario 2
May 23rd 2013 WIRN 2013 – Vietri sul Mare, Salerno, ItalyMay 23rd 2013
Conclusions and Future Works Fuzzy Decision Support System
identify potential impacts that can achieve one or more receptors through a set of migration paths
validated on Bt-maize1 and Brassica napus2 by the human experts of ISPRA
Future works Machine learning algorithms to learn the
FDSS knowledge base1 GM maize 2 GM colza