【摘要】 将不同训练数据子集和不同特征子集相结合,提出了一种基于小波变换特征提取的集成学习算法Wavelet-Forests.先随机划分特征集,用小波变换提取特征子集的特征,再用小波系数重构特征集训练基分类器.使用公认的WEKA平台验证了Wavelet-Forests算法的性能,与经典算法Bagging,AdaBoost和Random Forest相比,本文所提算法具有良好的泛化能力.更多还原

【Abstract】 To combine different subsets of training data and different feature subset,a new ensemble learning algorithm is put forward based on the feature extraction by wavelet transform.To create the training data for a base classifier,the feature set is randomly split into n subsets and wavelet is applied to each subset.The common platform WEKA has been used to validate the performance of Wavelet-Forests algorithm.The algorithm is superior to the classical algorithm Bagging,AdaBoost and Random Forest in comparison,and has good generalization ability.更多还原