Hidden Markov model

Ever wondered how you can predict outcomes without seeing all the details?

Imagine you're trying to guess how many candies are left in a jar without looking. You know the jar always contains a certain number of candies, but you can't see them all at once.

Think of the candies as a hidden Markov model where you can only see some candies (observations) influenced by the total number (hidden states). The Markov inequality helps you guess the minimum number of candies without checking every single one.

Example

If you know there are at least 100 candies and the average candy count per jar is 150, the Markov inequality suggests there's a high chance there are more than 150 candies.

Remember this

The Markov inequality helps you estimate the minimum number of candies in the jar based on average counts.

Text adapted from Wikipedia, licensed under CC BY-SA 4.0.

Related concepts

Markov's inequality

Ever wondered how math can predict unlikely events?

Chebyshev's inequality says: P(|X-μ| ≥ kσ) ≤ 1/k²

Ever wonder how likely it is to deviate from the average?

Hoeffding's inequality

Hoeffding's inequality bounds tail probability for sums of bounded random variables

Chebyshev's inequality

Chebyshev's inequality limits the probability of deviation from the mean

Cross-entropy H(p,q) = -Σ p(x) log q(x) measures how well q approximates p

Ever wondered how well we can guess the outcome of a random event?

Markov chain Monte Carlo

MCMC samples from complex posterior distributions

Swipe through 100 ML concepts daily

Open Pocket Polymath