A new machine learning method, "fractal learning", was proposed, which aims at acquiring knowledge from high-dimension and mass data in a better way. The research object of fractal learning is the self-similarity system. The inference of fractal learning is based on the study of the self-similarity of system. Fractal dimension used to describe the degree of the self-similarity of the system is the important parameter of fractal learning. The definition of the fractal learning is given and the basic principle of the fractal learning is explained along with the key technology during the learning process of the fractal learning. Finally, the method is applied to the reduction of the case base, which has achieved a good result%首先提出了一种新的机器学习方法,即分形学习,该方法旨在更好地从高维、海量数据中获取知识.分形论的研究对象为自相似系统,分形学习就是利用系统的自相似性来作为推理和学习的基本依据,其中分形维是定量描述系统自相似性的参数,也是应用于分形学习的重要参数；然后给出了分形学习的定义,阐述了其中的基本理论和学习过程中的关键技术；最后将该方法应用于案例库约简等方面的研究,取得了较好的效果.
展开▼
