【摘要】 传统的支持向量机(SVM)是一种有监督的机器学习方法,需要大量的有标签样本,而实际中对于有标签的样本数量十分有限且获得困难;直推式学习正是依据已知样本对特定的未知样本进行识别的方法与准则;研究了近年来直推式支持向量机学习算法及其改进算法,讨论了直推式学习算法的优缺点并对其发展进行了展望。更多还原

【Abstract】 Traditional support vector machine( SVM) is a kind of supervised machine learning algorithm and needs a lot of labeled samples,in practice,however,the number of the labeled samples is very limited and the labeled samples are difficult to obtain,as a result,transductive learning is the method and principle to distinguish unknown samples by depending on the known samples. This paper studies transductive support vector machine learning algorithm and its improved algorithm in recent years,discusses its advantages and disadvantages,and prospects its development.更多还原