Workshop on nonstationary models of pattern recognition and classifier combinations, Salamanca, Spain
28 March 2012 - 28 March 2012
Spain
A workshop on nonstationary models of pattern recognition and classifier combinations will take place from 28 to 30 March 2012 in Salamanca, Spain.
The progress of computer science has caused many entities to collect huge amounts of data, analysis of which is impossible by human beings. This is because simple methods often 'assume' that statistical properties of a discovered concept (which model is predicted) are unchanging. This has led to the rise of studies in an area called concept drift. Concept drift means that the statistical properties of the target variable, which the model is trying to predict, change over time in unforeseen ways. This can cause problems predictions to be less accurate over time.
The event will look at developing trends in pattern recognition which can exploit unique elementary classifier strengths and could adapt to changes in a classification model. Specific topics are set to include:
- adaptive and incremental learning;
- detecting concept change;
- theoretical foundations of multiple classifier systems;
- methods of classifier fusion;
- methods of decision making based on the information from different sources;
- methods of improving qualities of weak classifiers;
- methods of measuring and ensuring diversity in classifier ensembles;
- applications.For further information, please visit: http://hais.usal.es/?q=node/22(opens in new window)
The progress of computer science has caused many entities to collect huge amounts of data, analysis of which is impossible by human beings. This is because simple methods often 'assume' that statistical properties of a discovered concept (which model is predicted) are unchanging. This has led to the rise of studies in an area called concept drift. Concept drift means that the statistical properties of the target variable, which the model is trying to predict, change over time in unforeseen ways. This can cause problems predictions to be less accurate over time.
The event will look at developing trends in pattern recognition which can exploit unique elementary classifier strengths and could adapt to changes in a classification model. Specific topics are set to include:
- adaptive and incremental learning;
- detecting concept change;
- theoretical foundations of multiple classifier systems;
- methods of classifier fusion;
- methods of decision making based on the information from different sources;
- methods of improving qualities of weak classifiers;
- methods of measuring and ensuring diversity in classifier ensembles;
- applications.For further information, please visit: http://hais.usal.es/?q=node/22(opens in new window)