ass日本风韵熟妇pics男人扒开女人屁屁桶到爽|扒开胸露出奶头亲吻视频|邻居少妇的诱惑|人人妻在线播放|日日摸夜夜摸狠狠摸婷婷|制服 丝袜 人妻|激情熟妇中文字幕|看黄色欧美特一级|日本av人妻系列|高潮对白av,丰满岳妇乱熟妇之荡,日本丰满熟妇乱又伦,日韩欧美一区二区三区在线

The traditional support vector machines only deals with the binary classification. It has difficulty in solving the multi-class classification problem like the shift decision for the automatic transmission of the engineering vehicle. A shift decision algorithm that based on SVM-binary tree was presented. This method distributed classifier to nodes that constituted multi-class SVM. The number of SVM classifier and duplicated training samples could be reduced. The optimal shifting gear could be decided by the proposed approach, and the requirement of the engineering vehicle to the automatic shifting could be satisfied in time and accurately. The experiment showed that the support vector machines based on binary tree achieved better results than RBF neural network with genetics.