The package missMDA
The package missMDA is a companion to FactoMineR that permits to handle missing values in principal component methods (PCA, CA, MCA, MFA, FAMD). It performs single and multiple imputation.
Single imputation consists in replacing missing entries with plausible values. It leads to a complete data set that can be analyzed by any statistical methods.
missMDA imputes missing values in such a way that the imputed values have no weight (i.e. have no effect and the methods is performed with only the observed values) on the PCA (or any other methods) results.
Based on dimensionality reduction methods, the missMDA package successfully imputes large and complex datasets with quantitative and/or categorical variables. Indeed, it imputes data with PCA that take into account the similarities between the observations and the relationship between variables. It has proven to be very competitive in terms of quality of the prediction compared to the state of the art methods.
- missMDA handles missing values in:
- continuous data sets using the PCA model (See this video)
- categorical data sets using MCA (See this video)
- contingency table using CA
- mixed data using FAMD
- data set where variables are structured by groups using MFA
- missMDA generates multiple imputed data sets:
- for continuous data using the PCA model
- for categorical data using MCA
- missMDA allows you to visualize multiple imputation:
- in PCA
- in MCA
- missMDA gives hints to answer the question: Can you believe in the imputations? See this blog post.