Traj estimates a discrete mixture model for clustering of longitudinal data series. Groups may represent distinct subpopulations or alternatively, components of a discrete approximation for a potentially complex data distribution.

Supported distributions are: censored (or regular) normal, zero inflated (or regular) Poisson, and Bernoulli distributions (logistic model). The censored normal model is useful for psychometric scale data with censoring at a scale minimum and/or scale maximum, the zero inflated Poisson model useful for count data with more zeros than would be expected under the Poisson assumption, the Bernoulli model useful for 0/1 data, and the Beta model for continuous data bounded in the [0,1] interval. The model is appropriate for data with average values changing smoothly as a function of the dependent variable (time, age, ...). Some sharp changes can be handled through the inclusion of time dependent covariates.