Learning by Transduction

We describe a method for predicting a clas­ sification of an object given classifications of the objects in the training set, assuming that the pairs object/classification are generated by an i.i.d. process fr om a continuous proba­ bility distribution. Our method is a modifica­ tion of Vapnik's support-vector machine; its main novelty is that it gives not only the pre­ diction itself but also a practicable measure of the evidence found in support of that predic­ tion. We also describe a procedure for assign­ ing degrees of confidence to predictions made by the support vector machine. Some experi­ mental results are presented, and possible ex­ tensions of the algorithms are discussed.