Rolf Ergon

The paper corrects Hunt's model by using a tridiagonal covariance matrix instead of a diagonal one, improving step size estimation in evolutionary models.

This study compares directional random walk and weighted least squares methods for modeling sparse fossil data, finding that weighted least squares performs better in handling measurement errors.

This paper evaluates the directional random walk model for inferring evolutionary trends from sparse fossil data and finds that generalized least squares (GLS) is more reliable than GRW when measurement errors are large.

The paper analyzes fossil data to show that adaptive peak tracking models outperform other models in explaining trait changes over time.

The paper introduces a moving average smoothing method to extract obliquity cycles from fossil data, demonstrating improved trait predictions on stickleback fish fossils.

The paper simulates adaptive peak tracking in species over time using fossil data to infer environmental relationships and predict phenotypic changes.

The paper examines challenges in reaction norm predictions for wild populations, focusing on errors-in-variables, reference value uncertainty, and validation difficulties, and proposes a validation method using split data and prediction error analysis.