TY - JOUR
T1 - On the potential of ruled-based machine learning for disruption prediction on JET
AU - JET Contributors
AU - Lungaroni, M.
AU - Murari, A.
AU - Peluso, E.
AU - Vega, J.
AU - Farias, G.
AU - Gelfusa, M.
N1 - Publisher Copyright:
© 2018 University of Rome \”Tor Vergata\”.
PY - 2018/5
Y1 - 2018/5
N2 - In the last years, it has become apparent that detecting disruptions with sufficient anticipation time is an essential but not exclusive task of predictors. It is also important that the prediction is accompanied by appropriate qualifications of its reliability and it is formulated in mathematical terms appropriate for the task at hand (mitigation, avoidance, classification etc.). In this paper, a wide series of rule-based predictors, of the Classification and Regression Trees (CART) family, have been compared to assess their relative merits. An original refinement of the training, called noise-based ensembles, has allowed not only to obtain significantly better performance but also to increase the interpretability of the results. The final predictors can indeed be represented by a tree or a series of specific and clear rules. Such performance has been proved by analysing large databases of shots on JET with both the carbon wall and the ITER Like Wall. In terms of performance, the developed tools are therefore very competitive with other machine learning techniques, with the specificity of formulating the final models in terms of trees and simple rules.
AB - In the last years, it has become apparent that detecting disruptions with sufficient anticipation time is an essential but not exclusive task of predictors. It is also important that the prediction is accompanied by appropriate qualifications of its reliability and it is formulated in mathematical terms appropriate for the task at hand (mitigation, avoidance, classification etc.). In this paper, a wide series of rule-based predictors, of the Classification and Regression Trees (CART) family, have been compared to assess their relative merits. An original refinement of the training, called noise-based ensembles, has allowed not only to obtain significantly better performance but also to increase the interpretability of the results. The final predictors can indeed be represented by a tree or a series of specific and clear rules. Such performance has been proved by analysing large databases of shots on JET with both the carbon wall and the ITER Like Wall. In terms of performance, the developed tools are therefore very competitive with other machine learning techniques, with the specificity of formulating the final models in terms of trees and simple rules.
KW - Bagging
KW - Boosting
KW - Classification and regression trees
KW - Disruptions
KW - Machine learning predictors
KW - Noise-based ensembles
KW - Random forests
UR - http://www.scopus.com/inward/record.url?scp=85044164301&partnerID=8YFLogxK
U2 - 10.1016/j.fusengdes.2018.02.087
DO - 10.1016/j.fusengdes.2018.02.087
M3 - Article
AN - SCOPUS:85044164301
SN - 0920-3796
VL - 130
SP - 62
EP - 68
JO - Fusion Engineering and Design
JF - Fusion Engineering and Design
ER -