Influence functions

Influence functions are a statistical tool used to understand how small changes in a dataset, like adding or removing a data point, affect a model’s predictions or estimates. Think of them as a way to measure the impact of each individual data point on the overall model. This helps identify influential data points, diagnose model sensitivity, and improve robustness. Essentially, influence functions offer insight into which data points have the most effect on your model’s output without having to retrain the entire model each time.