Using stereo vision for autonomous mobile robot path-planning is a hot technology.The environment mapping and path-planning algorithms were introduced,and they
目前结合机器学习、文本分类的过滤方法成为研究热点。而这些过滤方法在邮件过滤时存在使用训练集数据量和表征数据的特征向量维数过多,引发"维灾难"和较大的运算量等问题。结合证据理论K近邻方法(evi-dence theory based K-nearest neighbors,EKNN)和直推式信度机(transductive confidence machines,TCM)算法思想,提出一种TCM-E