为了解决现有定量规律发现技术存在的发现范围狭窄、对噪声过于敏感、很难用于含干扰较大的实际观测数据等问题，将回归技术引入化简算法，提出了一个基于化简算法、原型回归和假说验证的多元函数规律发现的框架．根据该框架建立的知识发现系统，可用于发现蕴含于含有随机噪声的多变元观测数据中的数量规律 In order to solve the existing quantitative laws, we found that the technology has a narrow scope of discovery, is too sensitive to noise and difficult to be used in the actual observation data with large interference. The regression technique is introduced into the reduction algorithm, and a simplified algorithm based on simplification is proposed. Prototype regression and hypothesis verification of the multivariate rule of law found the framework. The knowledge discovery system based on this framework can be used to find the quantitative rules contained in multivariate observation data containing random noise