本文利用势函数的性质提出了一种简单而有效的新的势函数训练算法。与传统的势函数训练方法相比,该算法在很大程度上减少了学习的训练量,所得到的判决函数不仅在形式上更为简单,而且与训练样本集的边界有直接的映射关系。本文的工作有助于研究归纳学习的典型样本问题。 In this paper, a simple but effective new potential function training algorithm is proposed by using the properties of potential functions. Compared with the traditional method of potential function training, this algorithm greatly reduces the amount of learning training. The obtained decision function is not only simple in form, but also has a direct mapping relationship with the boundary of the training sample set. The work in this article helps to study the typical sample problem of inductive learning.