Arbiter meta-learning with dynamic selection of classifiers and its experimental investigation

In data mining, the selection of an appropriate classifier to estimate the value of an unknown attribute for a new instance has an essential impact to the quality of the classification result. Recently promising approaches using parallel and distributed computing have been presented. In this paper, we consider an approach that uses classifiers trained on a number of data subsets in parallel as in the arbiter meta-learning technique. We suggest that information is collected during the learning phase about the performance of the included base classifiers and arbiters and that this information is used during the application phase to select the best classifier dynamically. We evaluate our technique and compare it with the simple arbiter meta-learning using selected data sets from the UCI machine learning repository. The comparison results show that our dynamic meta-learning technique outperforms the arbiter meta-learning significantly in some cases but further profound analysis is needed to draw general conclusions.