# Enno Mammen

#### Professor for Mathematical Statistics

Heidelberg University

Institute for Applied Mathematics

MΛTHEMΛTIKON

Im Neuenheimer Feld 205

69120 Heidelberg, Germany

Phone: +49 (0) 6221 54 14180

Fax: +49 (0) 6221 54 14101

E-mail: mammen@math.uni-heidelberg.de

### Teaching SS 2021

### Research interests

At the start of my career I worked on asymptotic statistical decision theory (LeCam-Theory). In my thesis I solved a question raised by Lucien LeCam on the information contained in additional observations, published in an Annals of Statistics paper in 1989. Afterwards I studied parametric models with increasing dimension and nonparametric curve estimation problems. In particular, I looked on the performance of bootstrap methods in these models. In an influential paper in the Annals of Statistics from 1993, joint with Wolfgang Härdle, I proposed the wild bootstrap procedure for nonparametric regression problems and I studied wild bootstrap in another Annals paper of 1993 for high-dimensional linear models. Other work on bootstrap is contained in a Springer Lecture Notes in 1992. Further interests were nonparametric curve estimation under shape constraints (e.g. monotonicity or convexity), published among others in two Annals of Statistics papers in 1991. Furthermore I showed in an Annals paper from 1997 with O. Lepski and V. Spokoiny that one can achieve the same rates of convergence in Besov spaces by kernel smoothing compared with wavelet-estimators if one uses data-adaptive local bandwidths chosen with the Lepski-rule. With S. van de Geer I applied empirical process methods to study nonparametric estimators based on penalization. We wrote two Annals of Statistics papers in 1997. In one paper we used an L_{1}-penalty and showed that this leads to a sparse number of jumps of a function or its derivative, respectively. With A.B. Tsybakov I worked on estimation of sets in classification and discrimination. We summarized our research in two Annals papers (1995, 1999). Starting with Annals of Statistics in 1999, joint with O. Linton and J.P. Nielsen, on additive models I looked at structured nonparametric models where several nonparametric components enter into the modeling. In a series of papers I worked on this topic, in particular together with Oliver Linton, Cambridge; Joel Horowitz, Northwestern, Jens-Perch Nielsen, London, and Byeong Park, Seoul. Another more probabilistic interest, joint with Valentin Konakov, Moscow, were higher order limit statements of the convergence of Markov processes to diffusions.

### Education and career

### Selected publications in the last ten years

K. Gregory, E. Mammen and M. Wahl (2020) Optimal estimation of sparse high-dimensional additive models. *Annals of Statistics*, forthcoming.

C. Jentsch, E. Mammen and E. R. Lee (2020) Poisson reduced rank models with an application to political text data. *Biometrika*, forthcoming.

M. Hiabu, E. Mammen, M. D. Martínez Miranda, and J. P. Nielsen (2020) Smooth backfitting of proportional hazards with multiplicative components. *J. Amer. Statist. Assoc.*, forthcoming.

G. J. van den Berg, P. Bonev and E. Mammen (2020) Nonparametric instrumental variable methods for dynamic treatment evaluation. *The Review of Economics and Statistics* 102, 355 - 367.

A. Kreiß, E. Mammen and W. Polonik (2019) Nonparametric inference for continuous-time event counting and link-based dynamic network models. *Electronic Journal of Statistics*, 13, 2764 - 2829.

C. Breunig, E. Mammen and A. Simoni (2018) Nonparametric Estimation in case of Endogenous Selection. *J. Econometrics* 202(2), 268 - 285.

Y. K. Lee, E. Mammen, J. P. Nielsen and B. U. Park (2017) Operational time and in-sample density forecasting. *Ann. Statist.*, 45, 1312 - 1341.

E. R. Lee and E. Mammen (2016) Sparse High Dimensional Varying Coefficient Models. *Elect. J. Statistics*, 10, 855 - 894.

C. Conrad and E. Mammen (2016) Asymptotics for parametric GARCH-in-mean models. *J. of Econometrics*, 194, 319 - 329.

Y.K. Lee, E. Mammen, J. P. Nielsen and B.U. Park (2015) Asymptotics for In-Sample Density Forecasting. *Ann. Statist.*, 43 620 - 645.

V. Konakov, E. Mammen and J. Woerner (2014) Statistical convergence of Markov experiments to diffusion limits. *Bernoulli* 20, 623 - 644.

Y. K. Lee, E. Mammen and B. U. Park (2012) Flexible Generalized Varying Coefficient Regression Models. *Ann. Statist.* 40, 1906 - 1933.

E. Mammen, C. Rothe and M. Schienle (2012) Nonparametric Regression with Nonparametrically Generated Covariates. *Ann. Statist.*, 40, 1132 - 1170.

→ For all publications see:

**Publications and preprints, discussion of papers, book reviews**

### Major Contributions to early careeers of young researchers.

In the last ten years 11 students have finished a PhD thesis under my supervision. Three of my former PhD students are now full professors, three associate professors and two assistant professors. Topics of the PhD thesis include long memory GARCH models, nonparametric GARCH in mean models, additive nonparametric diffusion models, nonparametric regression tests, additive regression models with nonstationary covariates, nonparametric microeconometrics, local stationary nonparametric regression models, nonparametric instrumental regression, and high-dimensional nonparametric models. Currently I have 3 further PhD students, working on nonparametric Hawkes processes, on statistical theory of neural networks, and statistics of non-Euclidean data. Furthermore, I am co-supervising one PhD-student working on nonparametric survival analysis and in-sample forecasting.

Last edited: 2021-03-22 by jw