R Packages

Fused Lasso Additive Model [CRAN flam package]
Fits the fused lasso additive model. Feature selection is automatically performed, and each included feature’s fit is estimated to be piecewise constant with a small number of adaptively-chosen knots.
Reference: Petersen A, Witten D, and N Simon (2016). “Fused Lasso Additive Model.” Journal of Computational and Graphical Statistics, 25(4): 1005-1025. [arXiv]

Convex Regression With Interpretable Sharp Partitions [CRAN crisp package]
Implements convex regression with interpretable sharp partitions, which is an interpretable yet non-additive model. CRISP partitions the covariate space into blocks in a data-adaptive way, and fits a mean model within each block.
Reference: Petersen A, Simon N, and D Witten (2016). “Convex Regression with Interpretable Sharp Partitions.” Journal of Machine Learning Research, 17(94): 1-31. [pdf]

SCALPEL [CRAN scalpel package] [R vignette] [R vignette code] [R vignette video] [R vignette video code
Identifies the locations of neurons, and estimates their calcium concentrations over time using the SCALPEL method. Applicable to one- or two-photon calcium imaging data.
Reference: Petersen A, Simon N, and D Witten. “SCALPEL: Extracting Neurons from Calcium Imaging Data.” Annals of Applied Statistics, 12(4): 2430-2456. [pdf] [Section 6.2 supplementary] [Section 6.3.1 supplementary] [Section 6.3.2 supplementary] [Code and raw data]

Sparse Partially Linear Additive Trend Filtering [R splat package] [R splat manual]
Implements sparse partially linear additive trend filtering, which uses the data to determine which features to include in the model, whether to model each feature linearly or nonlinearly, and what form to use for the nonlinear functions.
Reference: Petersen A and D Witten (2019). “Data-Adaptive Additive Modeling.” Statistics in Medicine, 38(4): 583-600.

Shiny Apps

Fused Lasso Additive Model
Try out FLAM on [simulated data] or [your own data].

Selected apps developed for UW’s Biostat 536: Categorical Data Analysis
[Exploring Matched Case-Control Sampling]
[Collapsibility of Relative Risks, Risk Difference, and Odds Ratios]
[Understanding Linear and Quadratic Splines]