Fréquentiste inference is a method of statistical inference that emphasizes the frequency or proportion of data. In this framework, the parameters of a statistical model are considered fixed but unknown quantities. The primary goal is to use données obtenues à partir d'échantillons aléatoires pour tirer des conclusions sur ces paramètres.
In statistiques fréquentistes, the probability is interpreted as the long-run frequency of events occurring in repeated trials. For instance, if we say that a coin has a 50% probability of landing heads, we mean that if we flip the coin an infinite number of times, approximately half of the flips would result in heads.
Les concepts clés de l'inférence fréquentiste incluent :
- Estimations ponctuelles : Single values derived from sample data that serve as the best guess for a population parameter.
- Intervalles de confiance : A range of values that is likely to contain the population parameter with a specified level of confidence (e.g., 95%).
- Test d'hypothèse: A method for testing claims or hypotheses about population parameters. This involves formulating a null hypothesis, calculating a test statistic, and determining a p-value to assess the evidence against the null hypothesis.
Frequentist methods do not incorporate prior beliefs or evidence into the analysis. This distinguishes them from inférence bayésienne, which does consider prior information. Frequentist inference is widely used in various fields, including agriculture, medicine, and social sciences, due to its straightforward interpretation and application.