AIExplainer
Machine Learning Mathematics Intermediate 2 min read

What is log-odds?

A measure of the probability of an event occurring, calculated as the logarithm of the odds in favor of the event

Log-odds is a way to express the probability of something happening in a more convenient and mathematically useful form. It's calculated by taking the logarithm of the odds of an event, which is the ratio of the probability of the event happening to the probability of it not happening.

Think of log-odds like a thermometer for probabilities. Just as a thermometer shows the temperature, log-odds shows the 'temperature' of a probability, but on a scale that's easier to work with when doing calculations.

For example, in a medical diagnosis model, the log-odds of a patient having a certain disease might be calculated based on their symptoms and test results. A high log-odds score would indicate a high probability of the disease, while a low score would indicate a low probability.

Log-odds is used in machine learning and statistics to simplify calculations and make it easier to compare probabilities. It's particularly useful in logistic regression and other models where probabilities need to be combined or compared.

A common misconception is that log-odds is the same as probability, but it's actually a transformation of probability that makes it easier to work with in certain calculations.

The concept of log-odds has been around for centuries and was first used in the 18th century by mathematicians such as Pierre-Simon Laplace. It has since become a fundamental tool in statistics and machine learning.

logit log probability

Three products for different needs — explore what’s relevant to you.