The documentation page details FAMD (Factor Analysis for Mixed Data), a specialized principal component function from the popular R package FactoMineR. What is FAMD?
It is an exploratory multivariate method designed for datasets containing both continuous (quantitative) and categorical (qualitative) variables. You can think of it as a hybrid of Principal Component Analysis (PCA) and Multiple Correspondence Analysis (MCA):
For continuous variables: It acts like a PCA by scaling data to unit variance.
For categorical variables: It acts like an MCA by transforming categories into a disjunctive data table (crisp coding). Why use it?
Balanced Influence: It balances the math behind both variable types so they are on an equal footing to determine the dimensions of data variability.
Similarity Mapping: It reveals how similar different data observations (individuals) are based on highly mixed data profiles.
Relationship Visualization: It exposes how your continuous and categorical variables relate to each other. Key Outputs and Plots
The function generates comprehensive statistical tables and customizable graphical representations:
Individuals Map: A scatter plot mapping the similarities between data rows.
Correlation Circle: A classic circle diagram showing relationships among continuous variables.
Categories Map: A plot showing the associations between different categorical groupings.
If you are using this function, you can further enhance and customize these visualizations using the plot.FAMD documentation or the ggplot2-powered factoextra extension package.
Are you planning to run a mixed data analysis in R? If you want, tell me more about your dataset structure or what you want to achieve, and I can give you a quick sample script to get started!
FAMD: Factor Analysis for Mixed Data in FactoMineR – rdrr.io
Leave a Reply