ExploreModelMatrix

ExploreModelMatrix facilitates interactive exploration of design matrices for linear and generalized linear models to aid interpretation of model coefficients and formulation of linear contrasts for hypothesis testing.

Key Features:

• Interactive exploration: Visualizes how model coefficients are combined to produce fitted values for each combination of predictor variables, supporting extraction and interpretation of individual coefficients and linear contrasts.
• Visualization tools: Provides visual summaries that clarify relationships between input variables and their contributions to model predictions in linear and generalized linear models.
• Model diagnostics: Reports diagnostic measures for regular linear models associated with the design matrix, including variance inflation factors and the pseudoinverse of the design matrix to assess multicollinearity and model reliability.
• Built‑in model designs: Includes a collection of common linear model designs to illustrate and teach design-matrix structures and contrast specification.

Scientific Applications:

• Experimental design and interpretation: Assists in specifying contrasts and interpreting coefficient effects for designed experiments and observational studies.
• Hypothesis testing in linear models: Aids formulation and evaluation of linear contrasts for hypothesis tests in linear and generalized linear model frameworks.
• Applications in genetics, epidemiology, and ecology: Supports model specification, interpretation, and diagnostic assessment in domains that commonly use linear and generalized linear models.

Methodology:

The package takes a sample data table and a specified design formula to generate a design matrix using R's model.matrix() function and enables interactive inspection of how combinations of predictor variables influence fitted values.