gbmt: Group-Based Multivariate Trajectory Modeling Package

gbmt is an R package for estimating and analyzing group-based multivariate trajectory models, implementing the methodology from Nagin (2018) and Magrini (2022). The package provides an Expectation-Maximization algorithm that handles unbalanced panels and missing values, along with functionalities for prediction and graphical representation. It is authored by Alessandro Magrini and published under an Open Access (gold) license.

Use Cases

• Identifying latent subgroups in longitudinal data based on the implemented trajectory modeling.
• Predicting future outcomes for new subjects using the package's prediction functionality.
• Visualizing complex multivariate trajectories through the provided graphical representation tools.
• Analyzing longitudinal datasets with missing values or unbalanced panels using the EM algorithm.

Strengths

• Implements a published statistical methodology with citations to foundational papers (Nagin, 2018; Magrini, 2022).
• The algorithm supports unbalanced panel data and missing values, increasing applicability to real-world data.
• Provides multiple functionalities including estimation, prediction, and graphical representation in one package.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale applications.
• Last update date is unknown; freshness unverified.

Provenance

Source

Alessandro Magrini, Harvard University Press

Collection Method

Software package implementing a statistical modeling algorithm.