Error Rejection in Linearly Combined Multiple Classifiers
نویسندگان
چکیده
In this paper, the error-reject trade-off of linearly combined multiple classifiers is analysed in the framework of the minimum risk theory. Theoretical analysis described in [12,13] is extended for handling reject option and the optimality of the error-reject trade-off is analysed under the assumption of independence among the errors of the individual classifiers. Improvements of the error-reject trade-off obtained by linear classifier combination are quantified. Finally, a method for computing the coefficients of the linear combination and the value of the reject threshold is proposed. Experimental results on four different data sets are reported.
منابع مشابه
Analysis of Error-Reject Trade-off in Linearly Combined Classifiers
In this paper, a framework for the analysis of the error-reject trade-off in linearly combined classifiers is proposed. We start from a framework developed by Tumer and Ghosh [1,2]. We extend this framework and analyse some hypotheses under which the linear combination of classifier outputs can improve the error-reject trade-off of the individual classifiers. Experiments that support some of th...
متن کاملSimultaneous Monitoring of Multivariate Process Mean and Variability in the Presence of Measurement Error with Linearly Increasing Variance under Additive Covariate Model (RESEARCH NOTE)
In recent years, some researches have been done on simultaneous monitoring of multivariate process mean vector and covariance matrix. However, the effect of measurement error, which exists in many practical applications, on the performance of these control charts is not well studied. In this paper, the effect of measurement error with linearly increasing variance on the performance of ELR contr...
متن کاملAnalysis of error-reject trade-off in linearly combined multiple classifiers
In this paper, a theoretical and experimental analysis of the error-reject trade-off achievable by linearly combining the outputs of an ensemble of classifiers is presented. To this aim, the theoretical framework previously developed by Tumer and Ghosh for the analysis of the simple average rule without the reject option has been extended. Analytical results that allow to evaluate the improveme...
متن کاملBayesian Networks as Ensemble of Classifiers
Abstract Classification of real-world data poses a number of challenging problems. Mismatch between classifier models and true data distributions on one hand and the use of approximate inference methods on the other hand all contribute to inaccurate classification. Recent work on boosting by Schapire et al. and additive probabilistic models by Hastie et al. have shown that improved classificati...
متن کاملEnsemble Learning in Linearly Combined Classifiers Via Negative Correlation
We investigate the theoretical links between a regression ensemble and a linearly combined classification ensemble. First, we reformulate the Tumer & Ghosh model for linear combiners in a regression context; we then exploit this new formulation to generalise the concept of the “Ambiguity decomposition”, previously defined only for regression tasks, to classification problems. Finally, we propos...
متن کامل