Schätzung

Schätzung, in German, denotes estimation or an approximate value; in everyday language it often means a rough guess. In statistics and probability, Schätzung refers to the process of inferring an unknown population parameter from observed data or from a model based on data.

In statistical usage, there are two main concepts: point estimation and interval estimation. A point estimate is a single value proposed as an estimate of the parameter, produced by an estimator, which is a rule or function of the data. Interval estimation provides a range, such as a confidence interval, within which the parameter is believed to lie with a specified probability. Important properties of estimators include unbiasedness (on average equal to the true parameter), consistency (estimates converge to the true value with more data), and efficiency (low variance among unbiased estimators).

Common methods of estimation include Maximum-Likelihood-Schätzung (MLE), the method of least squares, and the method of moments. Bayesian estimation uses prior information to form a posterior distribution and derives estimators from that distribution. Estimation relies on assumptions about the data-generating process; model misspecification can lead to biased or misleading results. In applied contexts, estimation is central to fields such as econometrics, engineering, and the natural and social sciences, where parameters of models are inferred from sample data.

Aside from the technical sense, Schätzung also appears in everyday language as a rough guess or approximate value for a quantity when precise information is unavailable.