摘要
孤立点检测是数据挖掘中的一项广泛应用且较新的内容。根据孤立点的定义,提出一个基于偏离度的孤立点检测聚类分析算法。算法能够实现对异常数据进行处理。应用到学生成绩分析检测,保证实际聚类分析的准确性,数据的异常发现有利于学生教学的规划和推进。
The outliers detection algorithm is discussed, which is one of the widespread and new methods in data mining. Based on the definition of outliers, a clustering algorithm based on deviation degree for outlier detection is proposed. The algorithm can deal with abnormal data. The accuracy of data cluster analysis can be ensured by the application of the test of student achievement. Abnormal findings are conducive to the planning and promotion of student teaching.
作者
顾洪博
张继怀
GU Hong-bo;ZHANG Ji-huai(Northeast Petroleum University School of Computer ＆ Information Technology Heilongjiang Province Daqing 163318,China;DaQing City Ranghulu District Government DaQing City Heilongjiang Province 163712)
出处
《佳木斯大学学报（自然科学版）》
CAS
2018年第4期547-549,共3页
Journal of Jiamusi University：Natural Science Edition
基金
黑龙江省教育厅科研专项
东北石油大学引导性创新基金
关键词
孤立点检测
偏离度
聚类分析
outliers detection
deviation of distance-based
clustering analysis
