Code for reproducing simulations and analyses in "P-splines with an l1 penalty for repeated measures" by Segal, et al. (2018).

The accompanying R package is available at https://github.com/bdsegal/psplinesl1.

Contents

1. simulation: Code for the simulations.
   1. generate_data.R: Generate data for single group. The simulated dataset presented in the paper is provided in the psplinesl1 package as simData.
   2. generate_data2.R: Generate data for two groups. The simulated dataset presented in the paper is provided in the psplinesl1 package as simData2groups.
   3. plot_data.R: Plot the simulated datasets.
   4. l1_estimate.R: Fit the l1-penalized model to simData.
   5. l2_estimate.R: Fit the l2-penalized model to simData.
   6. bayes_esimate.R: Fit the Bayesian models. The stan code for fitting models with a laplace, normal, and diffuse prior on the finite order differences in coefficients is in bayes_lap.stan, bayes_norm.stan, and bayes_noPen.stan respectively.
   7. change_point_simulation_batch.R: Runs 100 simulations and saves the results in the batch folder. To conduct 1,000 simulations, run from the command line and pass in arguments 1-10.
   8. change_point_simulation_assess.R Plot the results of the change point simulation.
   9. l1_estimate_2groups.R: Fit the l1-penalized model to simData2groups.
   10. l2_estimate_2groups.R: Fit the l2-penalized model to simData2groups.
   11. coverage_prob_simulation.R: Simulate coverage probability and confidence interval width.
2. application: Code for the analysis of electrodermal activity (EDA) data collected as part of a stress study.
   1. process_data.R: pre-process raw data.
   2. analyze_EDA_l1.R: Fit the l1-penalized model.
   3. analyze_EDA_l2.R: Fit the l2-penalized model.
   4. analyze_EDA_l1_alt.R: Fit the l1-penalized model with an alternative correlation structure.
   5. analyze_EDA_l2_alt.R: Fit the l2-penalized model with an alternative correlation structure.
   6. bayes_norm: Folder containing files for fitting Bayesian model with normal prior on the "unpenalized" random effect coefficients. EDA_bayes_sqrt_lap_norm.R sets up and runs the model. The stan code is in EDA_lap_sqrt_norm.stan.
   7. bayes_cauchy: Folder containing files for fitting Bayesian model with Cauchy prior on the "unpenalized" random effect coefficients. EDA_bayes_sqrt_lap_cauchy.R sets up and runs the model. The stan code is in EDA_lap_sqrt_cauchy.stan.
   8. EDA_bayes_plots.R: Plot results from Bayesian models.
3. inference_knots_plots: Folder containing scripts to check the proof of observation 1 and make plots demonstrating our approach for approximate inference.
   1. observation1_check.R: Check proof of observation 1.
   2. approx_inference_plots.R: Plots demonstrating our approximate inference approach.

References

Segal, B. D., Elliott M., Braun T., and Jiang, H. (2018). P-splines with an l1 penalty for repeated measures. Electronic Journal of Statistics. 12(2), 3554-3600. doi.org/10.1214/18-EJS1487