bslme: Calculate test statistic using mixed effects model (LRT based...

In yishuwei2019/itlt: Interaction tree for longitudinal trajectories

Description

Mixed effects based model tests trajectories by testing model fitting. Trajectories are assumed to follow mixed effects models. Under the null hypothesis there are only one treatment effect curve for both nodes. Under the alternative, separate treatment effect curves are assumed for both nodes

Usage

bslme(D.traj, D.cov, tf, nknot = 5, p.value = 0.05, degree = 3)

Arguments

D.traj	dataset of purely longitudinal observations in wide format
D.cov	covariate matrix of intercept, treat, grp, treat * grp
tf	time frame
nknot	knot number
p.value	significance value required for a node to be splitted, default is 1
degree	degree of spline coefficients, default is 3

Value

(LRT) chi^2 test statistic of df(nknot) under the mixed effects model

yishuwei2019/itlt documentation built on July 8, 2021, 5:15 p.m.