Performance Evaluation of Missing-Value Imputation Clustering Based on a Multivariate Gaussian Mixture Model.

Area of Science:

• Machine Learning
• Data Science
• Statistical Modeling

Background:

• Clustering analysis is complicated by missing values in datasets.
• Existing methods for handling missing data in clustering can be suboptimal.

Purpose of the Study:

• To develop a dynamic clustering algorithm for datasets with missing values.
• To improve the accuracy and efficiency of clustering when data is incomplete.

Main Methods:

• A multivariate Gaussian mixture model was employed.
• An expectation-maximization algorithm was used for parameter and missing value estimation.
• Bayesian posterior probability was utilized for clustering individuals.

Main Results:

• The algorithm demonstrated fast convergence and accurate missing value imputation in simulations.
• Clustering accuracy was comparable to using complete datasets.
• The method outperformed deletion and mean imputation techniques in reducing misjudgment rates.

Conclusions:

• The proposed missing-value imputation clustering algorithm is effective and superior in specific scenarios.
• This approach offers a robust solution for clustering incomplete datasets.