How to update your beliefs using the math of posterior probability
In short: When you encounter new evidence, how should it change what you believe? Posterior probability is the mathematical engine for this update. By combining your prior knowledge with fresh data, you can calculate the revised likelihood of a hypothesis, turning raw observation into a precise, updated measure of uncertainty.
At its core, posterior probability is a conditional probability that emerges from applying Bayes' rule. It represents the state of knowledge regarding an uncertain proposition, such as a scientific hypothesis or specific parameter values, after incorporating both prior beliefs and a mathematical model of observed data. Essentially, it encapsulates everything known about a subject at a particular moment in time.
The process relies on a fundamental relationship: the posterior probability is proportional to the product of the likelihood and the prior probability. The likelihood function itself measures the probability of the evidence given specific parameters. To arrive at the final posterior, this product is divided by a normalizing constant, which ensures the resulting distribution is mathematically valid. While this framework is conceptually elegant, it is often not analytically tractable, requiring researchers to use numerical approximations to derive useful estimates.
Consider a school where 60% of students are boys and 40% are girls. If all boys wear trousers, but girls wear trousers or skirts with equal frequency, seeing a student in trousers provides specific information. While the prior probability of the student being a girl is 40%, the posterior probability—calculated after observing the trousers—drops to 25%. This demonstrates how new evidence shifts our understanding of a situation.
In fields like statistics and machine learning, posterior probabilities are vital for quantifying uncertainty. While some classification methods produce raw membership values, transforming these into probabilistic confidence levels allows for better post-processing and comparison. Because the posterior is conditioned on random data, it is itself a random variable, often summarized using credible intervals to provide a clearer picture of the remaining epistemic uncertainty.
Opens the Intellect+ feed with this card first. Swipe up for the next one.
Sources
- Summary by Intellect · source: Posterior probability
To cite this card, link to https://intellectplus.co.uk/app/science-posterior-probability.
How we source, check and correct cards: Editorial standards.
General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.
Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?
Related
More in Statistics and probability · More in Science · Latest · A–Z · All topics