arXiv subjects

Lasse Mohr

Publications and source records attributed to Lasse Mohr.

3 recordsLinked to original sources

arXiv2026

Births are difficult to predict even with rich survey and full-population register data

Major life events have proven difficult to predict. Does this reflect limits of theory, data, and algorithms, or the large role of chance? We examine one outcome - having a child within three years - through a near-ideal setting for prediction: a data challenge where 147 researchers predicted births for Dutch residents aged 18-45, using survey data and full-population registers. Methods ranged from logistic regression to a large language model and transformers. Predictions were moderately accurate (best F1: register 0.59, survey 0.76); advanced models did not outperform classical ones; and the larger registers did not beat the survey. Simulating the stochastic biology of conception and pregnancy, we estimated a predictive ceiling (survey F1 ~ 0.86-0.94, register 0.88-0.96). Observed performance falls short of this ceiling, implicating imperfect data, methods, and unmodelled chance, while the ceiling itself shows that chance in reproduction alone sets a non-trivial limit on predicting individual lives.

cs.LG↗

arXiv2025

The origins of large-scale structure in family networks

Family relations are the most fundamental of all social networks and encompass everyone. Family networks grow as individuals have children, creating connections between families, which over time create large and complex structures. While partner-choice homophily has been proposed as a key driver in this growth process, little is known about the connection between individual behavior and the emergent large-scale structure of family networks. Here, we analyze a unique population-complete family network, covering millions of individuals across several decades, enriched with demographic, educational, and geographic data from high-quality national registries. Drawing on the longitudinal coverage of our observations and using a series of growing-network models, we unravel how individual-level behavior shapes the large-scale network structure. Contrary to prevailing theories, we find that partner-choice homophily has little effect on the emergent large-scale structure. Instead, we identify two key drivers: First, partner-change behavior, where individuals leave one partner for another, creates `shortcuts' in the network akin to rewirings in the Watts-Strogatz model. These shortcuts decrease pathlengths and accelerate the emergence of meso-scale connected components. Second, we find that partner change is a self-exciting behavior, such that the probability of changing partner increases with an individual's prior number of partners. The self-exciting behavior accelerates the generation of large network components, with highly connected individuals functioning as network hubs. Accounting for this partner-change behavior, we are able to accurately capture multiple large-scale network properties of the empirical family network. Finally we show that homophily-driven behavior is not able to generate the observed network structure.

physics.soc-ph↗

arXiv2025

Long-Time and Short-Time Dynamics in a Weighted-Median Opinion Model on Networks

Social interactions influence people's opinions. In some situations, these interactions eventually yield a consensus opinion; in others, they can lead to opinion fragmentation and the formation of different opinion groups in the form of ``echo chambers''. Consider a social network of individuals with continuous-valued scalar opinions, and suppose that they can change their opinions when they interact with each other. In many models of the opinion dynamics of individuals in a network, it is common for opinion updates to depend on the mean opinion of interacting individuals. As an alternative, which may be more realistic in some situations, we study an opinion model with an opinion-update rule that depends on the weighted median of the opinions of interacting individuals. Through numerical simulations of our median-update opinion model, we investigate how the final opinion distribution depends on network structure. For configuration-model networks, we also derive a mean-field approximation for the asymptotic dynamics of the opinion distribution when there are infinitely many individuals. We numerically investigate its accuracy for short-time opinion dynamics on various networks.

physics.soc-ph↗