基于随机森林算法的CAM模板自动提取研究

Research on automatic CAM template extraction based on random forest clustering

将传统的数控编程与机器学习相结合,提出一种新的CAM模板的提取方法,利用随机森林算法分析数控工艺的关键参数,以随机森林相似性矩阵替代距离函数进行聚类分析,实现CAM模板自动提取。通过实际案例验证基于随机森林算法提取模板的可行性和准确性。结果表明:利用随机森林算法聚类分析得到的结果稳定性好,准确率达59.5%。

A novel approach to generate CAM template automatically was presented, which performed clustering based on random forest proximities instead of distance functions. On the basis of real database, a random forest classification was derived, and its feasibility and accuracy were identified. The experiment results indicated that the proposed approach was more stable and the accuracy was up to 59.5%.