改进极限学习机的移动界面模式半监督分类

Semi-supervised classification of mobile interface pattern using improved extreme learning machine

针对现有半监督分类方法无法对移动界面模式进行有效分类的问题,提出一种采用改进极限学习机的移动界面模式半监督分类方法。为了提高极限学习机的分类效果,利用改进的粒子群优化算法优化极限学习机的初始参数。根据移动界面模式数据的特点,利用主动学习和模糊C均值聚类提取信息丰富的未标记数据进行训练和标记。利用分类器实现对所有数据的分类。实验结果表明,该分类方法能够对移动界面模式数据进行有效和合理的分类。

Focused on the issue that the existing semi-supervised classification method cannot effectively classify mobile interface patterns, a semi-supervised classification of mobile interface pattern using improved extreme learning machine is proposed. Firstly, to enhance the classification effect of extreme learning machine, an improved particle swarm optimization algorithm is used to optimize the initial parameters of extreme learning machine. Secondly, according to the characteristics of mobile interface pattern data, active learning and fuzzy c-means clustering are employed to extract information rich unlabeled data for training and labeling. Finally, mobile interface pattern data are classified by using classifier. Experimental results show that the proposed semi-supervised classification method can classify the mobile interface pattern data effectively and reasonably.