Search

Your search keyword '"machine learning predictors"' showing total 11 results
Start Over Descriptor "machine learning predictors"
11 results on '"machine learning predictors"'

1. Machine learning models for dry matter and biomass estimates on cattle grazing systems.

Author

Defalque, Guilherme, Santos, Ricardo, Bungenstab, Davi, Echeverria, Diego, Dias, Alexandre, and Defalque, Cristiane

Subjects
*MACHINE learning, *EFFECT of environment on animals, *PEARSON correlation (Statistics), *GRAZING, *BIOMASS, *STANDARD deviations, *BIOMASS conversion
Abstract

[Display omitted] • Environmental and spectral data are commonly applied to estimate pasture parameters; • The dataset encompasses data collected in cattle grazing systems over two years; • Cattle characteristics have shown effects on biomass and dry matter predictions; • The models have attained better accuracy when organized into seasons of the year; • SVR and XGB models have achieved the best results on predicting the target variables; Monitoring pasture conditions contributes to the animals' decision-making process, avoiding supplementation losses, and improving cattle performance. Environmental parameters and herd characteristics can influence pasture quantity (biomass and dry matter) and controlling these parameters is a challenge nowadays. Some research work aim to monitor pasture quantity parameters from environmental and spectral data. Our research did not find studies considering the impact of the animals together with spectral data and environmental parameters to estimate pasture quantity parameters. This work aims to design machine models to estimate biomass and dry matter from Brachiaria brizantha cv. Marandu pastures, analyzing the effects of the environment and animals' features on the models. We collected pasture samples and analyzed them in a laboratory to determine the biomass and dry matter from 12 paddocks over two years in the spring and summer seasons. In addition, we acquired and associated spectral and environmental data and herd characteristics (gender and the number of animals), from the Nellore breed, with biomass and dry matter to design the models. Variable selection methods such as Pearson's correlation analysis and Recursive Feature Elimination (RFE) were applied to determine which parameters significantly affected the response variables. Non-linear and linear prediction models were designed and compared. The results indicated that the number of animals, gender, and vegetation indices have meaningful effects on the response variables. Furthermore, the non-linear models achieved expressive results predicting biomass and dry matter. The eXtreme Gradient Boosting - XGB Regressor (Adjusted R-Squared - R2 = 0.761 and Root Mean Squared Error - RMSE = 1914.80 kg/ha) achieved the best results in estimating biomass. The Support Vector Regressor (SVR) technique (Adjusted R2 = 0.773 and RMSE = 822.65 kg/ha) achieved the best results on estimating dry matter. To the best of our knowledge, this study is the first to consider cattle parameters, environmental and spectral data for estimating biomass and dry matter in grazing systems. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

2. On the potential of ruled-based machine learning for disruption prediction on JET.

Author

Lungaroni, M., Murari, A., Peluso, E., Vega, J., Farias, G., and Gelfusa, M.

Subjects
*MACHINE learning, *ELECTRIC inductance, *REGRESSION analysis, *DATA mining, *BINOMIAL distribution
Abstract

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. [ABSTRACT FROM AUTHOR]

Published
2018
Full Text
View/download PDF

3. Asynchronous and pathological windows of implantation: two causes of recurrent implantation failure.

Author

Sebastian-Leon, P., Garrido, N., Remohí, J., Pellicer, A., and Diaz-Gimeno, P.

Subjects
*EMBRYO transfer, *GENE expression, *REPRODUCTIVE technology, *PATIENTS, *GENOMICS, *COMPARATIVE studies, *ENDOMETRIUM, *INFERTILITY, *RESEARCH methodology, *MEDICAL cooperation, *RESEARCH, *EVALUATION research, *FETAL development, *RETROSPECTIVE studies, *GENE expression profiling
Abstract

Study Question: Is endometrial recurrent implantation failure (RIF) only a matter of an asynchronous (displaced) window of implantation (WOI), or could it also be a pathological (disrupted) WOI?Summary Answer: Our predictive results demonstrate that both displaced and disrupted WOIs exist and can present independently or together in the same RIF patient.What Is Known Already: Since 2002, many gene expression signatures associated with endometrial receptivity and RIF have been described. Endometrial transcriptomics prediction has been applied to the human WOI in two previous studies. One study describes endometrial RIF to be the result of a temporal displacement of the WOI. The other indicates that endometrial RIF can also result from a molecularly disrupted WOI without temporal displacement.Study Design, Size, Duration: Retrospective analysis was undertaken to compare WOI endometrial transcriptomics predictions in controls (n = 72) and RIF patients (n = 43). RIF was clinically designated by the absence of implantation after four or more transfers of high quality embryos or after the placement of 10 or more embryos in multiple transfers. Endometrial tissue samples were collected from LH + 5 to LH + 8. We compared the two molecular causes of RIF to signatures currently described in the literature. We propose a new transcriptomic RIF taxonomy to fill the gap between the two hypotheses and to guide the development of clinical detection and determination of both types of RIF.Participants/materials, Setting, Methods: Utilizing 115 gene expression profiles, two different predictive designs were developed: one considering RIF versus controls removing menstrual cycle timing, called the disrupted or pathological model, and another stratifying the WOI in transcriptomic profiles related to timing for predicting displacements. The predictive value of each model was compared between all signatures selected. We propose a new genomic approach that distinguishes between both types of RIF in the same sample cohort.Main Results and the Role Of Chance: From the 16 signatures analysed, we clearly predicted two causes of RIF-both a displaced WOI and an on-time but pathologically disrupted WOI. A high predictive value related to WOI profiles associated with menstrual cycle timing was found in most of the signatures. Specifically, 69% of the signatures analysed presented an accuracy higher than expected by chance in a range from 0.87 to 0.97. Displacements and disruptions were not molecularly independent, as some signatures were moderately associated with both causes. The gene and functional comparison between signatures revealed that they were not similar, although we did find functions in common and a cluster of moderate functional concordance between some of the signatures that predicted displacements (the highest Cohen's Kappa index were between 0.55 and 0.62 depending on the functional database). We propose a new transcriptomic RIF taxonomy to fill the gap between these prior studies and to establish methodology for detecting and distinguishing both types of RIF in clinical practice. Our findings indicate these two phenotypes could present independently or together in the same RIF patient. RIF patients designated by clinical criteria have been stratified transcriptomically as 18.6% with only a displaced WOI, 53.5% with a displaced and pathological WOI, 23.3% with only a disrupted WOI, and 4.7% could be a clinical RIF with non-endometrial origin. The new RIF transcriptomic taxonomy avoids menstrual cycle timing as a confounding variable that should be controlled for, distinguishing clearly between a disrupted and a displaced WOI for precision medicine in RIF.Limitations Reasons For Caution: The main objective of this study was to use transcriptomics to detect both RIF causes and to understand the role of transcriptomic signatures in these phenotypes. The predictive value in absolute terms for each signature was not indicative in these prediction designs; instead, the comparison between signatures was most important for prediction capability in the same sample cohort for both RIF causes. Clinical follow up of the RIF taxonomies proposed has not been analysed in this study, so further prospective clinical studies are necessary to determine the prevalence and penetrance of these phenotypes.Wider Implications Of the Findings: The main insight from this study is a new understanding of RIF taxonomy. Understanding how to classify RIF patients to distinguish clinically between a patient who could benefit from a personalized embryo transfer day and a patient with a disrupted WOI will enable identification and stratification for the research and development of new treatments. In addition, we demonstrate that basic research designs in endometrial transcriptomics cause masking of the study variable by the menstrual cycle timing.Study Funding/competing Interest(s): This research has been funded by IVI-RMA; the authors do not have any competing interests. [ABSTRACT FROM AUTHOR]

Published
2018
Full Text
View/download PDF

4. On the potential of ruled-based machine learning for disruption prediction on JET

Author

Stefan Matejcik, Francesco Romanelli, Augusto Pereira González, Jesús Vega, Bohdan Bieg, Emmanuele Peluso, Vladislav Plyusnin, José Vicente, Alberto Loarte, Michele Lungaroni, Bor Kos, Andrea Murari, Axel Jardin, Rajnikant Makwana, CHIARA MARCHETTO, Choong-Seock Chang, Manuel Garcia-munoz, Department of Physics, and Materials Physics

Subjects
Boosting (machine learning), Computer science, education, Machine learning, computer.software_genre, 114 Physical sciences, 01 natural sciences, Boosting, 010305 fluids & plasmas, Noise-based ensembles, Bagging, 0103 physical sciences, Classification and regression trees, General Materials Science, 010306 general physics, Civil and Structural Engineering, business.industry, Mechanical Engineering, Settore ING-IND/18 - Fisica dei Reattori Nucleari, Random forests, Random forest, Disruptions, Nuclear Energy and Engineering, Machine learning predictors, Artificial intelligence, business, computer
Abstract

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.

Published
2018
Full Text
View/download PDF

5. On the transfer of adaptive predictors between different devices for both mitigation and prevention of disruptions

Author

Michela Gelfusa, M. Lungaroni, A. Murari, Jesús Vega, Riccardo Rossi, G.A. Rattá, Emmanuele Peluso, and Jet Contributors, and Pasquale Gaudio

Subjects
Settore FIS/01, Nuclear and High Energy Physics, Adaptive strategies, ensembles of classifier, Computer science, Settore ING-IND/18 - Fisica dei Reattori Nucleari, adaptive learning, transfer learning, Condensed Matter Physics, disruptions, 01 natural sciences, Profit (economics), 010305 fluids & plasmas, Reliability engineering, Plasma current, Constant false alarm rate, mitigation, prevention, Software deployment, machine learning predictors, 0103 physical sciences, Adaptive learning, Overall performance, 010306 general physics, Transfer of learning
Abstract

Notwithstanding the efforts exerted over many years, disruptions remain a major impediment on the route to a magnetic confinement reactor of the Tokamak type. Machine learning predictors, relying on adaptive strategies, have recently proved to achieve very good performance. Even if their last generation implement a "from scratch" approach to learning, i.e. they can start predicting after the first example of each class (safe and disruptive), it would be extremely useful to profit from the experience of previous devices, when new machines come on online, to reduce excessive errors at the beginning of the learning process. In this paper, adaptive predictors, based on ensemble classifiers, have been operated on a series of AUG campaigns and then they have been deployed on several JET campaigns with the ILW, all together covering more than order of magnitude in plasma current. The criteria to normalise the signals and to translate the parameters of the predictors from one device to the other are discussed in detail. With regard to mitigation, the overall performance, both in terms of success rate and false alarms, are quite positive (98% success rate and only 1.9% false alarm rate). Encouraging results have also been obtained for prevention (94.2% success rate and only 7.7 % false alarm rate), by providing as inputs to the classifiers appropriate profile indicators. Even if they require significant refinements, adaptive predictors, capable of capitalising on the experience of smaller devices, have therefore become a serious candidate for deployment in the next generation of machines.

Published
2020
Full Text
View/download PDF

6. On the Potential of Adaptive Predictors and their Transfer between Different Devices for both Mitigation and Prevention of Disruptions

Author

Gelfusa M., Murari A., Rossi R., Lungaroni M., Peluso E., Rattà G., and Vega J.

Subjects
Tokamak, JET, Machine learning predictors, Mitigation and Prevention of Disruptions
Abstract

Notwithstanding the efforts exerted over many years, disruptions remain a major impediment on the route to a magnetic confinement reactor of the Tokamak type. Machine learning predictors, relying on adaptive strategies, have recently proved to achieve unprecedented performance on JET (with misclassifications of the order of a few per thousand both in terms of missed and false alarms) [1]. Such results are particularly relevant, because this last generation of adaptive predictors, based on ensemble classifiers, implement "from scratch" learning, i.e. they start predicting after the first example of each class (safe and disruptive) [2]. In order to show their potential to profit from the experience of previous devices when new machines come on online, specific adaptive predictors have also been operated on a series of AUG campaigns and then they have been deployed on several JET campaigns with the ITER Like Wall [3]. With regard to mitigation, the overall performance is extremely positive (errors of the order of a few per cent). Very encouraging results have also been obtained for disruption prevention. Adaptive predictors, capable of capitalising on the experience of smaller devices, have therefore become a serious candidate for deployment in the next generation of machines.

Published
2020

7. Stacking of predictors for the automatic classification of disruption types to optimise the control logic

Author

Andrea Murari, Riccardo Rossi, M. Baruzzo, Michela Gelfusa, and M. Lungaroni

Subjects
de-learning, Nuclear and High Energy Physics, Computer science, business.industry, Stacking, Settore ING-IND/18 - Fisica dei Reattori Nucleari, ensembles of classifiers, adaptive learning, transfer learning, Condensed Matter Physics, Machine learning, computer.software_genre, disruptions, Ensembles of classifiers, machine learning predictors, stacks of predictors, trajectory learning, Artificial intelligence, Adaptive learning, Trajectory learning, Control logic, Transfer of learning, business, computer
Abstract

Nowadays, disruption predictors, based on machine learning techniques, can perform well but they typically do not provide any information about the type of disruption and cannot predict the time remaining before the current quench. On the other hand, the automatic identification of the disruption type is a crucial aspect required to optimize the remedial actions and a prerequisite to forecasting the time left for intervening. In this work, a stack of machine learning tools is applied to the task of automatic classification of the disruption types. The strategy is implemented from scratch and completely adaptive; the predictors start operating after the first disruption and update their own models, following the evolution of the experimental program, without any human intervention. Moreover, they are designed to implement a form of transfer learning, in the sense that they identify autonomously the most important disruption classes, generating new ones when necessary. The results obtained are very encouraging in terms of both prediction performance and classification accuracy. On the other hand, regarding the narrowing of the warning times, some progress has been achieved, but new techniques will have to be devised to obtain fully satisfactory properties.

Published
2020
Full Text
View/download PDF

8. Adaptive learning for disruption prediction in non-stationary conditions

Author

Murari, A., Andersson Sundén, Erik, Binda, Federico, Cecconello, Marco, Conroy, Sean, Ericsson, Göran, Eriksson, Jacob, Hellesen, Carl, Hjalmarsson, Anders, Possnert, Göran, Primetzhofer, Daniel, Sahlberg, Arne, Sjöstrand, Henrik, Skiba, Mateusz, Weiszflog, Matthias, Zychor, I, Murari, A., Andersson Sundén, Erik, Binda, Federico, Cecconello, Marco, Conroy, Sean, Ericsson, Göran, Eriksson, Jacob, Hellesen, Carl, Hjalmarsson, Anders, Possnert, Göran, Primetzhofer, Daniel, Sahlberg, Arne, Sjöstrand, Henrik, Skiba, Mateusz, Weiszflog, Matthias, and Zychor, I

Abstract

For many years, machine learning tools have proved to be very powerful disruption predictors in tokamaks. On the other hand, the vast majority of the techniques deployed assume that the input data is independent and is sampled from exactly the same probability distribution for the training set, the test set and the final real time deployment. This hypothesis is certainly not verified in practice, since the experimental programmes evolve quite rapidly, resulting typically in ageing of the predictors and consequent suboptimal performance. This paper describes various adaptive training strategies that have been tested to maintain the performance of disruption predictors in non-stationary conditions. The proposed approaches have been implemented using new ensembles of classifiers, explicitly developed for the present application. The improvements in performance are unquestionable and, given the difficulties encountered so far in translating predictors from one device to another, the proposed adaptive methods from scratch can therefore be considered a useful option in the arsenal of alternatives envisaged for the next generation of devices, particularly at the very beginning of their operation., For complete list of authors see http://dx.doi.org/10.1088/1741-4326/ab1ecc

Published
2019
Full Text
View/download PDF

9. Adaptive predictors based on probabilistic SVM for real time disruption mitigation on JET

Author

Murari, A., Andersson Sundén, Erik, Binda, Federico, Cecconello, Marco, Conroy, Sean, Dzysiuk, Nataliia, Ericsson, Göran, Eriksson, Jacob, Hellesen, Carl, Hjalmarsson, Anders, Possnert, Göran, Sjöstrand, Henrik, Skiba, Mateusz, Weiszflog, Matthias, Zychor, I., Murari, A., Andersson Sundén, Erik, Binda, Federico, Cecconello, Marco, Conroy, Sean, Dzysiuk, Nataliia, Ericsson, Göran, Eriksson, Jacob, Hellesen, Carl, Hjalmarsson, Anders, Possnert, Göran, Sjöstrand, Henrik, Skiba, Mateusz, Weiszflog, Matthias, and Zychor, I.

Abstract

Detecting disruptions with sufficient anticipation time is essential to undertake any form of remedial strategy, mitigation or avoidance. Traditional predictors based on machine learning techniques can be very performing, if properly optimised, but do not provide a natural estimate of the quality of their outputs and they typically age very quickly. In this paper a new set of tools, based on probabilistic extensions of support vector machines (SVM), are introduced and applied for the first time to JET data. The probabilistic output constitutes a natural qualification of the prediction quality and provides additional flexibility. An adaptive training strategy 'from scratch' has also been devised, which allows preserving the performance even when the experimental conditions change significantly. Large JET databases of disruptions, covering entire campaigns and thousands of discharges, have been analysed, both for the case of the graphite and the ITER Like Wall. Performance significantly better than any previous predictor using adaptive training has been achieved, satisfying even the requirements of the next generation of devices. The adaptive approach to the training has also provided unique information about the evolution of the operational space. The fact that the developed tools give the probability of disruption improves the interpretability of the results, provides an estimate of the predictor quality and gives new insights into the physics. Moreover, the probabilistic treatment permits to insert more easily these classifiers into general decision support and control systems., For complete list of authors see http://dx.doi.org/10.1088/1741-4326/aaaf9c

Published
2018
Full Text
View/download PDF

10. Asynchronous and pathological windows of implantation: two causes of recurrent implantation failure

Author

Patricia Diaz-Gimeno, Patricia Sebastian-Leon, Antonio Pellicer, José Remohí, and Nicolás Garrido

Subjects
0301 basic medicine, Oncology, medicine.medical_specialty, Concordance, precision medicine, Biology, Endometrial tissue, 03 medical and health sciences, Endometrium, 0302 clinical medicine, Implantation failure, transcriptomic predictors, Internal medicine, medicine, polycyclic compounds, Humans, endometrial asynchrony, Embryo Implantation, endometrial pathology, endometrium, Pathological, Retrospective Studies, recurrent implantation failure, 030219 obstetrics & reproductive medicine, Gene Expression Profiling, Rehabilitation, Confounding, Obstetrics and Gynecology, Retrospective cohort study, Embryo Transfer, Penetrance, Gene expression profiling, gene expression signatures, 030104 developmental biology, Reproductive Medicine, machine learning predictors, window of implantation displacement, Female, Transcriptome, Infertility, Female, transcriptomic taxonomy
Abstract

STUDY QUESTION: Is endometrial recurrent implantation failure (RIF) only a matter of an asynchronous (displaced) window of implantation (WOI), or could it also be a pathological (disrupted) WOI? SUMMARY ANSWER: Our predictive results demonstrate that both displaced and disrupted WOIs exist and can present independently or together in the same RIF patient. WHAT IS KNOWN ALREADY: Since 2002, many gene expression signatures associated with endometrial receptivity and RIF have been described. Endometrial transcriptomics prediction has been applied to the human WOI in two previous studies. One study describes endometrial RIF to be the result of a temporal displacement of the WOI. The other indicates that endometrial RIF can also result from a molecularly disrupted WOI without temporal displacement. STUDY DESIGN, SIZE, DURATION: Retrospective analysis was undertaken to compare WOI endometrial transcriptomics predictions in controls (n = 72) and RIF patients (n = 43). RIF was clinically designated by the absence of implantation after four or more transfers of high quality embryos or after the placement of 10 or more embryos in multiple transfers. Endometrial tissue samples were collected from LH + 5 to LH + 8. We compared the two molecular causes of RIF to signatures currently described in the literature. We propose a new transcriptomic RIF taxonomy to fill the gap between the two hypotheses and to guide the development of clinical detection and determination of both types of RIF. PARTICIPANTS/MATERIALS, SETTING, METHODS: Utilizing 115 gene expression profiles, two different predictive designs were developed: one considering RIF versus controls removing menstrual cycle timing, called the disrupted or pathological model, and another stratifying the WOI in transcriptomic profiles related to timing for predicting displacements. The predictive value of each model was compared between all signatures selected. We propose a new genomic approach that distinguishes between both types of RIF in the same sample cohort. MAIN RESULTS AND THE ROLE OF CHANCE: From the 16 signatures analysed, we clearly predicted two causes of RIF-both a displaced WOI and an on-time but pathologically disrupted WOI. A high predictive value related to WOI profiles associated with menstrual cycle timing was found in most of the signatures. Specifically, 69% of the signatures analysed presented an accuracy higher than expected by chance in a range from 0.87 to 0.97. Displacements and disruptions were not molecularly independent, as some signatures were moderately associated with both causes. The gene and functional comparison between signatures revealed that they were not similar, although we did find functions in common and a cluster of moderate functional concordance between some of the signatures that predicted displacements (the highest Cohen's Kappa index were between 0.55 and 0.62 depending on the functional database). We propose a new transcriptomic RIF taxonomy to fill the gap between these prior studies and to establish methodology for detecting and distinguishing both types of RIF in clinical practice. Our findings indicate these two phenotypes could present independently or together in the same RIF patient. RIF patients designated by clinical criteria have been stratified transcriptomically as 18.6% with only a displaced WOI, 53.5% with a displaced and pathological WOI, 23.3% with only a disrupted WOI, and 4.7% could be a clinical RIF with non-endometrial origin. The new RIF transcriptomic taxonomy avoids menstrual cycle timing as a confounding variable that should be controlled for, distinguishing clearly between a disrupted and a displaced WOI for precision medicine in RIF. LIMITATIONS REASONS FOR CAUTION: The main objective of this study was to use transcriptomics to detect both RIF causes and to understand the role of transcriptomic signatures in these phenotypes. The predictive value in absolute terms for each signature was not indicative in these prediction designs; instead, the comparison between signatures was most important for prediction capability in the same sample cohort for both RIF causes. Clinical follow up of the RIF taxonomies proposed has not been analysed in this study, so further prospective clinical studies are necessary to determine the prevalence and penetrance of these phenotypes. WIDER IMPLICATIONS OF THE FINDINGS: The main insight from this study is a new understanding of RIF taxonomy. Understanding how to classify RIF patients to distinguish clinically between a patient who could benefit from a personalized embryo transfer day and a patient with a disrupted WOI will enable identification and stratification for the research and development of new treatments. In addition, we demonstrate that basic research designs in endometrial transcriptomics cause masking of the study variable by the menstrual cycle timing. STUDY FUNDING/COMPETING INTEREST(S): This research has been funded by IVI-RMA; the authors do not have any competing interests.

Published
2018

11. On the potential of ruled-based machine learning for disruption prediction on JET

Author

Lungaroni, M., Murari, A., Peluso, E., Vega, J., Farias, G., Gelfusa, M., Abduallev, S., Abhangi, M., Abreu, P., Afzal, M., Aggarwal, K. M., Ahlgren, T., Ahn, J. H., Aho-Mantila, L., Aiba, N., Airila, M., Albanese, R., Aldred, V., Alegre, D., Alessi, E., Aleynikov, P., Alfier, A., Alkseev, A., Allinson, M., Alper, B., Alves, E., Ambrosino, G., Ambrosino, R., Amicucci, L., Amosov, V., Sunden, E. Andersson, Angelone, M., Anghel, M., Angioni, C., Appel, L., Appelbee, C., Arena, P., Ariola, M., Arnichand, H., Arshad, S., Ash, A., Ashikawa, N., Aslanyan, V., Asunta, O., Auriemma, F., Austin, Y., Avotina, L., Axton, M. D., Ayres, C., Bacharis, M., Baciero, A., Baiao, D., Bailey, S., Baker, A., Balboa, I., Balden, M., Balshaw, N., Bament, R., Banks, J. W., Baranov, Y. F., Barnard, M. A., Barnes, D., Barnes, M., Barnsley, R., Wiechec, A. Baron, Orte, L. Barrera, Baruzzo, M., Basiuk, V., Bassan, M., Bastow, R., Batista, A., Batistoni, P., Baughan, R., Bauvir, B., Baylor, L., Bazylev, B., Beal, J., Beaumont, P. S., Beckers, M., Beckett, B., Becoulet, A., Bekris, N., Beldishevski, M., Bell, K., Belli, F., Bellinger, M., Belonohy, E., Ben Ayed, N., Benterman, N. A., Bergsaker, H., Bernardo, J., Bernert, M., Berry, M., Bertalot, L., Besliu, C., Beurskens, M., Bieg, B., Bielecki, J., Biewer, T., Bigi, M., Bilkova, P., Binda, F., Bisoffi, A., Bizarro, J. P. S., Bjorkas, C., Blackburn, J., Blackman, K., Blackman, T. R., Blanchard, P., Blatchford, P., Bobkov, V., Boboc, A., Bodnar, G., Bogar, O., Bolshakova, I., Bolzonella, T., Bonanomi, N., Bonelli, F., Boom, J., Booth, J., Borba, D., Borodin, D., Borodkina, I., Botrugno, A., Bottereau, C., Boulting, P., Bourdelle, C., Bowden, M., Bower, C., Bowman, C., Boyce, T., Boyd, C., Boyer, H. J., Bradshaw, J. M. A., Braic, V., Bravanec, R., Breizman, B., Bremond, S., Brennan, P. D., Breton, S., Brett, A., Brezinsek, S., Bright, M. D. J., Brix, M., Broeckx, W., Brombin, M., Broslawski, A., Brown, D. P. D., Brown, M., Bruno, E., Bucalossi, J., Buch, J., Buchanan, J., Buckley, M. A., Budny, R., Bufferand, H., Bulman, M., Bulmer, N., Bunting, P., Buratti, P., Burckhart, A., Buscarino, A., Busse, A., Butler, N. K., Bykov, I., Byrne, J., Cahyna, P., Calabro, G., Calvo, I., Camenen, Y., Camp, P., Campling, D. C., Cane, J., Cannas, B., Capel, A. J., Card, P. J., Cardinali, A., Carman, P., Carr, M., Carralero, D., Carraro, L., Carvalho, B. B., Carvalho, I., Carvalho, P., Casson, F. J., Castaldo, C., Catarino, N., Caumont, J., Causa, F., Cavazzana, R., Cave-Ayland, K., Cavinato, M., Cecconello, M., Ceccuzzi, S., Cecil, E., Cenedese, A., Cesario, R., Challis, C. D., Chandler, M., Chandra, D., Chang, C. S., Chankin, A., Chapman, I. T., Chapman, S. C., Chernyshova, M., Chitarin, G., Ciraolo, G., Ciric, D., Citrin, J., Clairet, F., Clark, E., Clark, M., Clarkson, R., Clatworthy, D., Clements, C., Cleverly, M., Coad, J. P., Coates, P. A., Cobalt, A., Coccorese, V., Cocilovo, V., Coda, S., Coelho, R., Coenen, J. W., Coffey, I., Colas, L., Collins, S., Conka, D., Conroy, S., Conway, N., Coombs, D., Cooper, D., Cooper, S. R., Corradino, C., Corre, Y., Corrigan, G., Cortes, S., Coster, D., Couchman, A. S., Cox, M. P., Craciunescu, T., Cramp, S., Craven, R., Crisanti, F., Croci, G., Croft, D., Crombe, K., Crowe, R., Cruz, N., Cseh, G., Cufar, A., Cullen, A., Curuia, M., Czarnecka, A., Dabirikhah, H., Dalgliesh, P., Dalley, S., Dankowski, J., Darrow, D., Davies, O., Davis, W., Day, C., Day, I. E., De Bock, M., de Castro, A., de la Cal, E., de la Luna, E., De Masi, G., de Pablos, J. L., De Temmerman, G., De Tommasi, G., de Vries, P., Deakin, K., Deane, J., Agostini, F. Degli, Dejarnac, R., Delabie, E., den Harder, N., Dendy, R. O., Denis, J., Denner, P., Devaux, S., Devynck, P., Di Maio, F., Di Siena, A., Di Troia, C., Dinca, P., D'Inca, R., Ding, B., Dittmar, T., Doerk, H., Doerner, R. P., Donne, T., Dorling, S. E., Dormido-Canto, S., Doswon, S., Douai, D., Doyle, P. T., Drenik, A., Drewelow, P., Drews, P., Duckworth, Ph., Dumont, R., Dumortier, P., Dunai, D., Dunne, M., Duran, I., Durodie, F., Dutta, P., Duval, B. P., Dux, R., Dylst, K., Dzysiuk, N., Edappala, P. V., Edmond, J., Edwards, A. M., Edwards, J., Eich, Th., Ekedahl, A., El-Jorf, R., Elsmore, C. G., Enachescu, M., Ericsson, G., Eriksson, F., Eriksson, J., Eriksson, L. G., Esposito, B., Esquembri, S., Esser, H. G., Esteve, D., Evans, B., Evans, G. E., Evison, G., Ewart, G. D., Fagan, D., Faitsch, M., Falie, D., Fanni, A., Fasoli, A., Faustin, J. M., Fawlk, N., Fazendeiro, L., Fedorczak, N., Felton, R. C., Fenton, K., Fernades, A., Fernandes, H., Ferreira, J., Fessey, J. A., Fevrier, O., Ficker, O., Field, A., Fietz, S., Figueiredo, A., Figueiredo, J., Fil, A., Finburg, P., Firdaouss, M., Fischer, U., Fittill, L., Fitzgerald, M., Flammini, D., Flanagan, J., Fleming, C., Flinders, K., Fonnesu, N., Fontdecaba, J. M., Formisano, A., Forsythe, L., Fortuna, L., Fortuna-Zalesna, E., Fortune, M., Foster, S., Franke, T., Franklin, T., Frasca, M., Frassinetti, L., Freisinger, M., Fresa, R., Frigione, D., Fuchs, V., Fuller, D., Futatani, S., Fyvie, J., Gal, K., Galassi, D., Galazka, K., Galdon-Quiroga, J., Gallagher, J., Gallart, D., Galvao, R., Gao, X., Gao, Y., Garcia, J., Garcia-Carrasco, A., Garcia-Munoz, M., Gardarein, J. -L., Garzotti, L., Gaudio, P., Gauthier, E., Gear, D. F., Gee, S. J., Geiger, B., Gerasimov, S., Gervasini, G., Gethins, M., Ghani, Z., Ghate, M., Gherendi, M., Giacalone, J. C., Giacomelli, L., Gibson, C. S., Giegerich, T., Gil, C., Gil, L., Gilligan, S., Gin, D., Giovannozzi, E., Girardo, J. B., Giroud, C., Giruzzi, G., Gloeggler, S., Godwin, J., Goff, J., Gohil, P., Goloborod'ko, V., Gomes, R., Goncalves, B., Goniche, M., Goodliffe, M., Goodyear, A., Gorini, G., Gosk, M., Goulding, R., Goussarov, A., Gowland, R., Graham, B., Graham, M. E., Graves, J. P., Grazier, N., Grazier, P., Green, N. R., Greuner, H., Grierson, B., Griph, F. S., Grisolia, C., Grist, D., Groth, M., Grove, R., Grundy, C. N., Grzonka, J., Guard, D., Guerard, C., Guillemaut, C., Guirlet, R., Gurl, C., Utoh, H. H., Hackett, L. J., Hacquin, S., Hagar, A., Hager, R., Hakola, A., Halitovs, M., Hall, S. J., Cook, S. P. Hallworth, Hamlyn-Harris, C., Hammond, K., Harrington, C., Harrison, J., Harting, D., Hasenbeck, F., Hatano, Y., Hatch, D. R., Haupt, T. D. V., Hawes, J., Hawkes, N. C., Hawkins, J., Hawkins, P., Haydon, P. W., Hayter, N., Hazel, S., Heesterman, P. J. L., Heinola, K., Hellesen, C., Hellsten, T., Helou, W., Hemming, O. N., Hender, T. C., Henderson, M., Henderson, S. S., Henriques, R., Hepple, D., Hermon, G., Hertout, P., Hidalgo, C., Highcock, E. G., Hill, M., Hillairet, J., Hillesheim, J., Hillis, D., Hizanidis, K., Hjalmarsson, A., Hobirk, J., Hodille, E., Hogben, C. H. A., Hogeweij, G. M. D., Hollingsworth, A., Hollis, S., Homfray, D. A., Horacek, J., Hornung, G., Horton, A. R., Horton, L. D., Horvath, L., Hotchin, S. P., Hough, M. R., Howarth, P. J., Hubbard, A., Huber, A., Huber, V., Huddleston, T. M., Hughes, M., Huijsmans, G. T. A., Hunter, C. L., Huynh, P., Hynes, A. M., Iglesias, D., Imazawa, N., Imbeaux, F., Imrisek, M., Incelli, M., Innocente, P., Irishkin, M., Ivanova-Stanik, I., Jachmich, S., Jacobsen, A. S., Jacquet, P., Jansons, J., Jardin, A., Jarvinen, A., Jaulmes, F., Jednorog, S., Jenkins, I., Jeong, C., Jepu, I., Joffrin, E., Johnson, R., Johnson, T., Johnston, Jane, Joita, L., Jones, G., Jones, T. T. C., Hoshino, K. K., Kallenbach, A., Kamiya, K., Kaniewski, J., Kantor, A., Kappatou, A., Karhunen, J., Karkinsky, D., Karnowska, I., Kaufman, M., Kaveney, G., Kazakov, Y., Kazantzidis, V., Keeling, D. L., Keenan, T., Keep, J., Kempenaars, M., Kennedy, C., Kenny, D., Kent, J., Kent, O. N., Khilkevich, E., Kim, H. T., Kim, H. S., Kinch, A., King, C., King, D., King, R. F., Kinna, D. J., Kiptily, V., Kirk, A., Kirov, K., Kirschner, A., Kizane, G., Klepper, C., Klix, A., Knight, P., Knipe, S. J., Knott, S., Kobuchi, T., Koechl, F., Kocsis, G., Kodeli, I., Kogan, L., Kogut, D., Koivuranta, S., Kominis, Y., Koeppen, M., Kos, B., Koskela, T., Koslowski, H. R., Koubiti, M., Kovari, M., Kowalska-Strzeciwilk, E., Krasilnikov, A., Krasilnikov, V., Krawczyk, N., Kresina, M., Krieger, K., Krivska, A., Kruezi, U., Ksiazek, I., Kukushkin, A., Kundu, A., Kurki-Suonio, T., Kwak, S., Kwiatkowski, R., Kwon, O. J., Laguardia, L., Lahtinen, A., Laing, A., Lam, N., Lambertz, H. T., Lane, C., Lang, P. T., Lanthaler, S., Lapins, J., Lasa, A., Last, J. R., Laszynska, E., Lawless, R., Lawson, A., Lawson, K. D., Lazaros, A., Lazzaro, E., Leddy, J., Lee, S., Lefebvre, X., Leggate, H. J., Lehmann, J., Lehnen, M., Leichtle, D., Leichuer, P., Leipold, F., Lengar, I., Lennholm, M., Lerche, E., Lescinskis, A., Lesnoj, S., Letellier, E., Leyland, M., Leysen, W., Li, L., Liang, Y., Likonen, J., Linke, J., Linsmeier, Ch., Lipschultz, B., Liu, G., Liu, Y., Lo Schiavo, V. P., Loarer, T., Loarte, A., Lobel, R. C., Lomanowski, B., Lomas, P. J., Lonnroth, J., Lopez, J. M., Lopez-Razola, J., Lorenzini, R., Losada, U., Lovell, J. J., Loving, A. B., Lowry, C., Luce, T., Lucock, R. M. A., Lukin, A., Luna, C., Lungu, C. P., Lungu, M., Lunniss, A., Lupelli, I., Lyssoivan, A., Macdonald, N., Macheta, P., Maczewa, K., Magesh, B., Maget, P., Maggi, C., Maier, H., Mailloux, J., Makkonen, T., Makwana, R., Malaquias, A., Malizia, A., Manas, P., Manning, A., Manso, M. E., Mantica, P., Mantsinen, M., Manzanares, A., Maquet, Ph., Marandet, Y., Marcenko, N., Marchetto, C., Marchuk, O., Marinelli, M., Marinucci, M., Markovic, T., Marocco, D., Marot, L., Marren, C. A., Marshal, R., Martin, A., Martin, Y., Martin de Aguilera, A., Martinez, F. J., Martin-Solis, J. R., Martynova, Y., Maruyama, S., Masiello, A., Maslov, M., Matejcik, S., Mattei, M., Matthews, G. F., Maviglia, F., Mayer, M., Mayoral, M. L., May-Smith, T., Mazon, D., Mazzotta, C., McAdams, R., McCarthy, P. J., McClements, K. G., McCormack, O., McCullen, P. A., McDonald, D., McIntosh, S., McKean, R., McKehon, J., Meadows, R. C., Meakins, A., Medina, F., Medland, M., Medley, S., Meigh, S., Meigs, A. G., Meisl, G., Meitner, S., Meneses, L., Menmuir, S., Mergia, K., Merrigan, I. R., Mertens, Ph., Meshchaninov, S., Messiaen, A., Meyer, H., Mianowski, S., Michling, R., Middleton-Gear, D., Miettunen, J., Militello, F., Militello-Asp, E., Miloshevsky, G., Mink, F., Minucci, S., Miyoshi, Y., Mlynar, J., Molina, D., Monakhov, I., Moneti, M., Mooney, R., Moradi, S., Mordijck, S., Moreira, L., Moreno, R., Moro, F., Morris, A. W., Morris, J., Moser, L., Mosher, S., Moulton, D., Muraro, A., Murphy, S., Asakura, N. N., Na, Y. S., Nabais, F., Naish, R., Nakano, T., Nardon, E., Naulin, V., Nave, M. F. F., Nedzelski, I., Nemtsev, G., Nespoli, F., Neto, A., Neu, R., Neverov, V. S., Newman, M., Nicholls, K. J., Nicolas, T., Nielsen, A. H., Nielsen, P., Nilsson, E., Nishijima, D., Noble, C., Nocente, M., Nodwell, D., Nordlund, K., Nordman, H., Nouailletas, R., Nunes, I., Oberkofler, M., Odupitan, T., Ogawa, M. T., O'Gorman, T., Okabayashi, M., Olney, R., Omolayo, O., O'Mullane, M., Ongena, J., Orsitto, F., Orszagh, J., Oswuigwe, B. I., Otin, R., Owen, A., Paccagnella, R., Pace, N., Pacella, D., Packer, L. W., Page, A., Pajuste, E., Palazzo, S., Pamela, S., Panja, S., Papp, P., Paprok, R., Parail, V., Park, M., Diaz, F. Parra, Parsons, M., Pasqualotto, R., Patel, A., Pathak, S., Paton, D., Patten, H., Pau, A., Pawelec, E., Soldan, C. Paz, Peackoc, A., Pearson, I. J., Pehkonen, S. -P., Penot, C., Pereira, A., Pereira, R., Puglia, P. P. Pereira, von Thun, C. Perez, Peruzzo, S., Peschanyi, S., Peterka, M., Petersson, P., Petravich, G., Petre, A., Petrella, N., Petrzilka, V., Peysson, Y., Pfefferle, D., Philipps, V., Pillon, M., Pintsuk, G., Piovesan, P., Pires dos Reis, A., Piron, L., Pironti, A., Pisano, F., Pitts, R., Pizzo, F., Plyusnin, V., Pomaro, N., Pompilian, O. G., Pool, P. J., Popovichev, S., Porfiri, M. T., Porosnicu, C., Porton, M., Possnert, G., Potzel, S., Powell, T., Pozzi, J., Prajapati, V., Prakash, R., Prestopino, G., Price, D., Price, M., Price, R., Prior, P., Proudfoot, R., Pucella, G., Puglia, P., Puiatti, M. E., Pulley, D., Purahoo, K., Puetterich, Th., Rachlew, E., Rack, M., Ragona, R., Rainford, M. S. J., Rakha, A., Ramogida, G., Ranjan, S., Rapson, C. J., Rasmussen, J. J., Rathod, K., Ratta, G., Ratynskaia, S., Ravera, G., Rayner, C., Rebai, M., Reece, D., Reed, A., Refy, D., Regan, B., Regana, J., Reich, M., Reid, N., Reimold, F., Reinhart, M., Reinke, M., Reiser, D., Rendell, D., Reux, C., Reyes Cortes, S. D. A., Reynolds, S., Riccardo, V., Richardson, N., Riddle, K., Rigamonti, D., Rimini, F. G., Risner, J., Riva, M., Roach, C., Robins, R. J., Robinson, S. A., Robinson, T., Robson, D. W., Roccella, R., Rodionov, R., Rodrigues, P., Rodriguez, J., Rohde, V., Romanelli, F., Romanelli, M., Romanelli, S., Romazanov, J., Rowe, S., Rubel, M., Rubinacci, G., Rubino, G., Ruchko, L., Ruiz, M., Ruset, C., Rzadkiewicz, J., Saarelma, S., Sabot, R., Safi, E., Sagar, P., Saibene, G., Saint-Laurent, F., Salewski, M., Salmi, A., Salmon, R., Salzedas, F., Samaddar, D., Samm, U., Sandiford, D., Santa, P., Santala, M. I. K., Santos, B., Santucci, A., Sartori, F., Sartori, R., Sauter, O., Scannell, R., Schlummer, T., Schmid, K., Schmidt, V., Schmuck, S., Schneider, M., Schoepf, K., Schworer, D., Scott, S. D., Sergienko, G., Sertoli, M., Shabbir, A., Sharapov, S. E., Shaw, A., Shaw, R., Sheikh, H., Shepherd, A., Shevelev, A., Shumack, A., Sias, G., Sibbald, M., Sieglin, B., Silburn, S., Silva, A., Silva, C., Simmons, P. A., Simpson, J., Simpson-Hutchinson, J., Sinha, A., Sipila, S. K., Sips, A. C. C., Siren, P., Sirinelli, A., Sjostrand, H., Skiba, M., Skilton, R., Slabkowska, K., Slade, B., Smith, N., Smith, P. G., Smith, R., Smith, T. J., Smithies, M., Snoj, L., Soare, S., Solano, E. R., Somers, A., Sommariva, C., Sonato, P., Sopplesa, A., Sousa, J., Sozzi, C., Spagnolo, S., Spelzini, T., Spineanu, F., Stables, G., Stamatelatos, I., Stamp, M. F., Staniec, P., Stankunas, G., Stan-Sion, C., Stead, M. J., Stefanikova, E., Stepanov, I., Stephen, A. V., Stephen, M., Stevens, A., Stevens, B. D., Strachan, J., Strand, P., Strauss, H. R., Strom, P., Stubbs, G., Studholme, W., Subba, F., Summers, H. P., Svensson, J., Swiderski, L., Szabolics, T., Szawlowski, M., Szepesi, G., Suzuki, T. T., Tal, B., Tala, T., Talbot, A. R., Talebzadeh, S., Taliercio, C., Tamain, P., Tame, C., Tang, W., Tardocchi, M., Taroni, L., Taylor, D., Taylor, K. A., Tegnered, D., Telesca, G., Teplova, N., Terranova, D., Testa, D., Tholerus, E., Thomas, J., Thomas, J. D., Thomas, P., Thompson, A., Thompson, C. -A., Thompson, V. K., Thorne, L., Thornton, A., Thrysoe, A. S., Tigwell, P. A., Tipton, N., Tiseanu, I., Tojo, H., Tokitani, M., Tolias, P., Tomes, M., Tonner, P., Towndrow, M., Trimble, P., Tripsky, M., Tsalas, M., Tsavalas, P., Jun, D. Tskhakaya, Turner, I., Turner, M. M., Turnyanskiy, M., Tvalashvili, G., Tyrrell, S. G. J., Uccello, A., Ul-Abidin, Z., Uljanovs, J., Ulyatt, D., Urano, H., Uytdenhouwen, I., Vadgama, A. P., Valcarcel, D., Valentinuzzi, M., Valisa, M., Olivares, P. Vallejos, Valovic, M., Van De Mortel, M., Van Eester, D., Van Renterghem, W., van Rooij, G. J., Varje, J., Varoutis, S., Vartanian, S., Vasava, K., Vasilopoulou, T., Verdoolaege, G., Verhoeven, R., Verona, C., Rinati, G. Verona, Veshchev, E., Vianello, N., Vicente, J., Viezzer, E., Villari, S., Villone, F., Vincenzi, P., Vinyar, I., Viola, B., Vitins, A., Vizvary, Z., Vlad, M., Voitsekhovitch, I., Vondracek, P., Vora, N., Vu, T., Pires de Sa, W. W., Wakeling, B., Waldon, C. W. F., Walkden, N., Walker, M., Walker, R., Walsh, M., Wang, E., Wang, N., Warder, S., Warren, R. J., Waterhouse, J., Watkins, N. W., Watts, C., Wauters, T., Weckmann, A., Weiland, J., Weisen, H., Weiszflog, M., Wellstood, C., West, A. T., Wheatley, M. R., Whetham, S., Whitehead, A. M., Whitehead, B. D., Widdowson, A. M., Wiesen, S., Wilkinson, J., Williams, J., Williams, M., Wilson, A. R., Wilson, D. J., Wilson, H. R., Wilson, J., Wischmeier, M., Withenshaw, G., Withycombe, A., Witts, D. M., Wood, D., Wood, R., Woodley, C., Wray, S., Wright, J., Wright, J. C., Wu, J., Wukitch, S., Wynn, A., Xu, T., Yadikin, D., Yanling, W., Yao, L., Yavorskij, V., Yoo, M. G., Young, C., Young, D., Young, I. D., Young, R., Zacks, J., Zagorski, R., Zaitsev, F. S., Zanino, R., Zarins, A., Zastrow, K. D., Zerbini, M., Zhang, W., Zhou, Y., Zilli, E., Zoita, V., Zoletnik, S., and Zychor, I.

Subjects
classification and regression trees, random forests, noise-based ensembles, machine learning predictors, boosting, disruptions, bagging
Abstract

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.

Catalog

Books, media, physical & digital resources
See catalog results