LASSO type penalized spline regression for binary data

Abstract Background Generalized linear mixed models (GLMMs), typically used for analyzing correlated data, can also be smoothing by considering the knot coefficients from a regression spline as random effects. The resulting are called semiparametric (SPMMs). Allowing to follow normal distribution with mean zero and constant variance is equivalent using penalized ridge type penalty. We introduce least absolute shrinkage selection operator (LASSO) penalty in SPMM setting at knots Laplace double exponential zero. Methods adopt Bayesian approach use Markov Chain Monte Carlo (MCMC) algorithm model fitting. Through simulations, we compare performance of curve fitting LASSO that binary data. apply proposed method obtain smooth curves data on relationship between amount pack years smoking risk developing chronic obstructive pulmonary disease (COPD). Results performs well simple shapes association outperforms when shape complex or linear. Conclusion demonstrated captured dose-response better than Ridge SPMM.