一种基于极限学习机的缺失数据填充方法

A METHOD FOR MISSING DATA IMPUTATION BASED ON EXTREME LEARNING MACHINE

数据处理过程中经常会遇到不完备数据需要填充的问题,寻求简单有效的缺失数据填充方法非常重要。针对该情况,提出一种基于极限学习机ELM(Extreme Learning Machine)的缺失数据填充方法,通过极限学习机网络建模,建立需要填充的缺失属性与其他属性的非线性映射模型。实验结果表明:该方法具有非常好的填充效果。

In data processing process the problems of having to impute incomplete data are often encountered,so it is important to look for a simple and effective missing data imputation method. In view of this,the paper presents an extreme learning machine-based method for missing data imputation. Based on extreme learning machine modelling it builds a nonlinear mapping model of missing attributes with the need of imputation as well as other attributes. Experimental result shows that the new algorithm has excellent performance in imputation.