An interactive tool that calculates posterior probabilities using log‑space arithmetic for numerical stability. Ideal for statisticians, data scientists, and researchers who need fast, reliable Bayesian updates.
Try the CalculatorHandles extreme likelihoods and priors without underflow or overflow.
Instant posterior updates as you type.
Ensures only valid numeric log values are accepted.
Download results as CSV or JSON with one click.
Optimized for mobile, tablet, and desktop screens.
Switch themes for comfortable viewing in any lighting.
Provide the log‑likelihood and log‑prior values for your hypothesis.
The tool computes log‑evidence using the LogSum trick for numerical stability.
Posterior = exp(logLikelihood + logPrior – logEvidence).
Log‑space arithmetic transforms multiplication into addition and avoids underflow/overflow when dealing with very small or very large probabilities. This yields numerically stable posterior calculations.
Enter the natural logarithm of the likelihood and prior (logLikelihood and logPrior). They can be any real number; positive values correspond to likelihoods/priors >1, negative to <1.
We compute log‑evidence using the LogSum trick: logEvidence = LogSum(logLikelihood + logPrior, log(1‑exp(logLikelihood)) + logComplement). Then logPosterior = logLikelihood + logPrior – logEvidence, and posterior = exp(logPosterior).
No. All calculations are performed client‑side in your browser. No data leaves your device.
Experience fast, reliable Bayesian updates with a single click.