一种基于k-means聚类技术的快速选择性Bagging Trees集成算法研究

A Study on Quick Selective Bagging Trees Ensemble Algorithm Based on K-means Cluster Technology

选择性集成算法是目前机器学习关注的热点之一。在对一海藻繁殖案例研究的基础上,提出了一种基于k-means聚类技术的快速选择性Bagging Trees集成算法;同时与传统统计方法和一些常用的机器学习方法相比较,发现该算法具有较小的模型推广误差和更高的预测精度的优点,而且其运行的效率也得到了较大的提高。

Selective ensemble algorithm now becomes a hot topic in machine learning. In this paper, based on a case study of algae propagation, the authors draw a new ensemble algorithm, a quick selective bagging trees ensemble algorithm based on k - means cluster technology. And contrasted with the traditional statistical methods and some machine learning methods, this new algorithm proposed in this paper is not only a model to promote small generalized error and the higher accuracy, but also costs much little time than other algorithms.