Boosting with Prior for Accurate Classification

Adaptive Boosting (AdaBoost) based meta learning algorithms generate an accurate classifier ensemble using a algorithm with only moderate accuracy guarantees. These have been designed to work in typical supervised settings and hence use labeled training data along base form ensemble. However, significant knowledge about the solution space might be available data. The convergence rate of AdaBoost improved such knowledge. An effective way incorporate into boosting is presented this paper. Using several synthetic real datasets, empirical evidence reported show effectiveness proposed method.Significant improvements obtained by applying method for detecting roads aerial images.