Data Imputation Methods for Missing Values in the Context of Clustering

Abstract

Missing data is a common problem for data clustering quality. Most real-life datasets have missing data, which in turn has some effect on clustering tasks. This chapter investigates the appropriate data treatment methods for varying missing data scarcity distributions including gamma, Gaussian, and beta distributions. The analyzed data imputation methods include mean, hot-deck, regression, k-nearest neighbor, expectation maximization, and multiple imputation. To reveal the proper methods to deal with missing data, data mining tasks such as clustering is utilized for evaluation. With the experimental studies, this chapter identifies the correlation between missing data imputation methods and missing data distributions for clustering tasks. The results of the experiments indicated that expectation maximization and k-nearest neighbor methods provide best results for varying missing data scarcity distributions.

Occupational resource profiles for an addressee orientation in occupational health management: a segmentation analysis

“…Frequently, some individuals showed up to five missing values, so listwise case exclusion would have resulted in a sample of n = 794. For this reason, k -nearest neighbor imputation ( k = 10) of the missing data was performed, which was used in similar contexts and was able to provide good results for different distributions of missing data (Liao et al, 2014 ; Cleophas and Zwinderman, 2016 ; Aktaş et al, 2019 ). In k -nearest neighbor imputation, the mean values of the neighboring values are formed and assigned to the missing value (von der Hude, 2020 ).…”