基于相近原则的半指导直推学习机

TRANSDUCTIVE LEARNING MACHINES BASED ON THE AFFINITY-RULE FOR SEMI-SUPERVISED PROBLEM

机器学习研究中的一个重要课题是如何有效结合已标识数据和未标识数据去推断未标识点标识.本文利用相近原则和直推方法求解这个问题.源于直观事实的相近原则是指"在输入空间中相近的对象其输出也相近".为求得满足这个原则的半指导学习问题的解,我们给出了作为一般方法的基于相近原则的半指导问题直推学习机.得到解的解析表达和迭代算法,公式形式简洁,计算实现容易.给出实例验证该方法在解决实际问题中的有效性,并用图例与支撑向量机和半指导支撑向量机的解作了对比.

One of central problems in machine learning is how to effectively combine labeled with unlabeled data to infer labels for the unlabeled data. In this paper, the affinity-rule and transductive method are used. The affinity-rule which comes from the intuitive fact is that two objects in input space are close then their outputs are also close. To get the solution of semi-supervised learning problem satisfying the affinity-rule, transductive learning machines based on the affinity-rule is introduced as a general method. The analytic solution for the problem and its iterated algorithm are obtained. Its results are simple in formula and easy for calculation. Some examples about pattern classification are given to show our method practical validity, and our cutlines are compared with ones of SVM and S^3VM.