Abstract: In this paper，we propose the orbits generated lattice of learning subspace in Lie-group Machine Learning（LML） and its corresponding basic conceptions，which include sample set in Lie-group machine learning，orbits generated lattices theory and algorithm.Synchronously，this paper analyzes the example and compares its results with those of the C4.5 decision tree learning algorithm，from which we find our algorithm is much superior to C4.5 in validity of categorization，thus it can be believed that the theory is feasible and the algorithm is valid.

Key words: Lie-group Machine Learning（LML）, learning subspace, orbits generated lattice

摘要： 给出了李群机器学习（LML）的学习子空间轨道生成格及相关的基本概念，包括：李群机器学习中的样例数据集，轨道生成格理论及其算法，同时也给出了实例验证分析，并与决策树学习算法C4.5作比较，在分类的正确性方面优于C4.5算法，由此进一步证明了该理论的可行性以及算法的有效性。

关键词: 李群机器学习, 学习子空间, 轨道生成格