1: \begin{abstract}

2:   We propose a unified framework for

3:   likelihood-based regression modeling when the response variable has finite

4:   support. Our work is motivated by the fact that, in practice, observed data

5:   are discrete and bounded. The proposed methods assume a model which includes

6:   models previously considered for interval-censored variables with log-concave

7:   distributions as special cases. The resulting log-likelihood is concave, which

8:   we use to establish asymptotic normality of its maximizer as the number of

9:   observations $n$ tends to infinity with the number of parameters $d$ fixed,

10:   and rates of convergence of $L_1$-regularized estimators when the true

11:   parameter vector is sparse and $d$ and $n$ both tend to infinity with $\log(d)

12:   / n \to 0$. We consider an inexact proximal Newton algorithm for computing

13:   estimates and give theoretical guarantees for its convergence. The range of

14:   possible applications is wide, including but not limited to survival analysis

15:   in discrete time, the modeling of outcomes on scored surveys and

16:   questionnaires, and, more generally, interval-censored regression. The

17:   applicability and usefulness of the proposed methods are illustrated in

18:   simulations and data examples.

19: \end{abstract}

20: