基于聚类分析的缺失数据最近邻填补算法

摘要

Data missing exists in various research fields universally and the missed data will cause serious impact on computational performance and effect.In order to improve the accuracy of missing data filling,we propose a cluster analysis-based nearest neighbour filling algorithm for the missing data.After analysing the cluster data,the algorithm assigns the weights according to the categories;moreover,it improves the calculation method and the calculation means of filling value based on the MGNN (Mahalanobis-gray and nearest neighbour) algorithm.Experimental results show that the filling accuracy of the method is higher than the traditional KNN algorithm and MGNN algorithm.%数据缺失在各个研究领域中普遍存在，缺失的数据会对计算的性能与结果产生严重的影响。为提高填补缺失数据的准确度，提出一种基于聚类分析的缺失数据最近邻填补算法。该算法在对数据聚类分析后根据类别分配权重，在MGNN（Mahalanobis-Gray and Nearest Neighbor）算法的基础上改进了计算方法和填充值的计算方式。实验结果表明，该方法填补的准确度比传统KNN和MGN N算法要高。