Handling missing data in clustering using multiple imputation - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Handling missing data in clustering using multiple imputation

Résumé

Multiple imputation techniques are often used for addressing the missing data issue in statistical analysis. It is presented how it can be considered for addressing missing values in the context of clustering. To achieve this goal, a novel imputation method is presented entitled FCS-homo, as well as a pooling method for the set of partitions obtained from each imputed data set. The proposed methodology is evaluated using a simulation study in comparison with state-of-the-art methods. It started by treating the case where the observations are generated from a Gaussian mixture model with missing random values. Experiments are based on various real data sets where the distribution of the variables is unknown. These first results tend to show that multiple imputation is an efficient method for handling missing data in clustering, especially when the data distribution is unknown.
Fichier non déposé

Dates et versions

hal-04529644 , version 1 (02-04-2024)

Identifiants

  • HAL Id : hal-04529644 , version 1

Citer

Vincent Audigier, Ndèye Niang. Handling missing data in clustering using multiple imputation. 16th International Conference of the ERCIM WG on Computational and Methodological Statistics, Dec 2023, Berlin (Germany), Germany. ⟨hal-04529644⟩
3 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More