Adaboost classification of TJ-II Thomson Scattering images

Gonzalo Farias, Sebastián Dormido-Canto, Jesús Vega, Ismael Martínez, Luis Alfaro, Francisco Martínez

Research output: Contribution to journalArticlepeer-review

7 Scopus citations


Experiments in thermonuclear fusion generate thousands of signals. Machine learning techniques have shown to be very suitable for implementing pattern recognition systems to fusion databases. The huge amount of data involves performing the analysis in high-dimensional spaces. This makes difficult the searching of patterns with similar properties, which normally produces overfitting. During last years, the use of boosting algorithms has become very popular for avoiding overfitting and building generalized data-driven models. Boosting allows achieving a highly accurate, robust and fast classification by combining many relatively simple rules. In this work, we propose the use of Adaboost algorithm to classify Thomson Scattering images of the TJ-II fusion device. The results are compared with previous works on similar databases.

Original languageEnglish
Pages (from-to)759-763
Number of pages5
JournalFusion Engineering and Design
StatePublished - Nov 2017


  • Adaboost
  • Image classifier
  • TJ-II
  • Thomson Scattering diagnostic


Dive into the research topics of 'Adaboost classification of TJ-II Thomson Scattering images'. Together they form a unique fingerprint.

Cite this