Mathematics
Bayesian Statistics
Quick fact
Bayes' theorem was discovered by Thomas Bayes in the 18th century, but only gained widespread use in the 20th century when computers made complex calculations practical.
Why this is interesting
You guessed there was a 50% chance of rain, then you saw dark clouds. How should you change your guess? Bayes' rule tells you exactly how to update your beliefs with new evidence.
Read the full explanation
Understanding Bayesian Statistics
Imagine you're holding a coin. Before you flip it, you might think it's fair (a 50% chance of heads). But after seeing 10 heads in a row, you start to suspect something is off. Bayesian statistics formalizes this intuition: you start with a prior probability—your initial belief. Then you collect data—the coin flips. Finally, you combine the prior and data using Bayes' theorem to get a posterior probability—your updated belief. The beauty is that the posterior can serve as a new prior for future updates, making learning continuous and coherent.
A deeper explanation
At the heart of Bayesian statistics is Bayes' theorem, expressed as P(H|E) = P(E|H) P(H) / P(E). Here, P(H) is the prior, E is the evidence, P(E|H) is the likelihood (how likely the evidence is if the hypothesis is true), and P(E) is the total probability of the evidence. The theorem shows how to invert conditional probabilities: it tells you how the evidence should change your belief in the hypothesis. This framework treats probability as a subjective degree of belief rather than a long-run frequency, which allows for straightforward incorporation of expert knowledge. It also automatically handles uncertainty, because the posterior is a full distribution, not just a point estimate. This makes Bayesian methods powerful for decision-making and for updating predictions as new data arrives, which is essential in fields like medicine, finance, and AI.