Computer Engineering and Applications ›› 2015, Vol. 51 ›› Issue (12): 111-117.
Previous Articles Next Articles
XIU Yu1, WANG Jun2, WANG Zhongqun1, HUANG Subin1
Online:
Published:
修 宇1,王 骏2,王忠群1,皇苏斌1
Abstract: Aiming at the problem of high computational complexity of Greed Max-Cut Graph semi-supervised learning algorithm(GGMC), an improved Greed Max-Cut Graph semi-supervised learning algorithm based on Early stopping strategy, called GGMC-Estop, is proposed. According to the experimental analysis on that object function value in optimization procedure of GGMC, the algorithm is improved in which two early stopping strategies are applied to stop GGMC training and prediction. Standard propagation is used to predict the label of data over the whole graph in one step. Experimental results on typical data sets show that the computational amount using the improved algorithm is far less than that of using GGMC algorithm, while the performance of classification for these two algorithms is almost approximate.
Key words: graph semi-supervised learning, greedy max-cut, early stopping strategy, object function value
摘要: 针对贪心最大割图半监督学习算法(简称GGMC)计算复杂度较高的问题,提出一种改进的贪心最大割图半监督学习算法(简称GGMC-Estop)。依据对GGMC算法优化过程中目标函数变化趋势的实验分析,采取两种在迭代初期停止GGMC算法运行策略,继而通过一次标准的标签传播步骤预测图上所有样本的标记来实施对GGMC的改进。典型数据集的仿真实验结果表明,在取得相近分类性能的同时,改进算法在计算速度上有很大的提高。
关键词: 图半监督学习, 贪心最大割, 早期停止策略, 目标函数值
XIU Yu1, WANG Jun2, WANG Zhongqun1, HUANG Subin1. Experimental analysis on object function value of GGMC algorithm and algorithm improvement[J]. Computer Engineering and Applications, 2015, 51(12): 111-117.
修 宇1,王 骏2,王忠群1,皇苏斌1. GGMC算法目标函数值实验分析与算法改进[J]. 计算机工程与应用, 2015, 51(12): 111-117.
0 / Recommend
Add to citation manager EndNote|Ris|BibTeX
URL: http://cea.ceaj.org/EN/
http://cea.ceaj.org/EN/Y2015/V51/I12/111