Lennart Oelschläger

GitHub Email Google Scholar

Hi, I’m Lennart! 👋🙂

I am currently a research associate in the Econometrics Group at Bielefeld University and recently completed my PhD. I like data science and math, teaching both, and developing open source software. In my spare time, you can find me in front of a chessboard or above the clouds.

NoteResearch Interests

• Discrete choice models, in particular preference heterogeneity and Bayesian estimation of latent class mixed multinomial probit models.
• Numerical optimization, in particular methods for efficient and robust initialization.
• Hidden Markov models, in particular hierarchical extensions to improve trend detection in financial time series across different time scales.

NoteRecent Talks

• Faster and more robust maximum likelihood estimation for random utility models, ICMC in Puerto Varas, April 1, 2024
• Modeling unobserved choice behavior heterogeneity, 13th YRW at Bielefeld University, February 16, 2024
• Facilitating probit likelihood optimization, Statistische Woche in Dortmund, September 12, 2023
• Bayesian probit models for preference classification, IWSM in Dortmund, July 18, 2023
• Initialization of numerical optimization in R, 12th YRW at Bielefeld University, September 7, 2023
• Advances in the initialization of probit model estimation and the {ino} R package, 11th YRW at Bielefeld University, November 18, 2022

NoteTeaching

Several lectures and tutorials in statistics and math. See here for a complete list.

NoteSoftware

ao	Alternating Optimization
choicedata	Working with Choice Data
fHMM	Fitting Hidden Markov Models to Financial Data
ino	Initialization of Numerical Optimization
normalize	Centering and Scaling of Numeric Data
oeli	Some Utilities for Developing Data Science Software
optimizeR	Unified Framework for Numerical Optimizers
portion	Extracting a Data Portion
RprobitB	Bayesian Probit Choice Modeling
trackopt	Track Numerical Optimization
vntrs	Variable Neighborhood Trust Region Search