基于改进的随机森林的缺失值填补
首发时间:2018-12-05
摘要:缺失值数据对于自然数据研究领域是一个常见的数据异常,该异常极大影响了对数据样本理解。由于目前研究多基于单一领域数据进行缺失值填补,泛用性较差。因此,本文针对通用数据的缺失值问题进行研究,基于随机森林算法通过动态调整参数以适应不同领域数据,并提高数据准确性。实验结果表明,基于改进的随机森林的缺失值填补具有较高的准确度和泛用性。
For information in English, please click here
Missing Value Filling Based on Improved Random Forests
Abstract:Missing value data is a common abnormal data in the field of natural data research, which greatly affects the understanding of data samples. Since the current research is based on single-domain data for missing value filling, the versatility is poor. Therefore, this paper studies the missing value of general data, based on random forest algorithm to dynamically adjust parameters to adapt to different domain data, and improve data accuracy. The result of experiments shows thatmissing value filling based on improved random forest has higher accuracy and versatility.
Keywords: Missing value General data Random Forest Dynamic parameter
基金:
引用
No.****
同行评议
勘误表
基于改进的随机森林的缺失值填补
评论
全部评论0/1000