Bayesian Thinking

Photo by Ben Wicks on Unsplash

You’re trying to figure out if a friend is telling the truth. You already have a certain belief about their honesty (your prior belief). Then, your friend tells you a story that seems a bit fishy (the data).

Bayesian thinking is updating your beliefs based on this new information. If the story doesn’t fit with what you know about your friend, you might become less convinced of their honesty. But if the story lines up with their character, you might become even more confident.

Outline:

1. So, Bayesian Thinking.
2. Main Elements.
3. Bayes’ Theorem
4. Example.
5. But why do we need this?

So, Bayesian Thinking:

Bayesian thinking is a statistical approach that uses probability to update beliefs as new evidence becomes available.

Main Elements:

1. Prior Probability: This is the initial belief or estimate of the probability of an event occurring. It’s based on existing knowledge or assumptions.
2. Likelihood: This is the probability of observing the data given a particular hypothesis. It’s how well the data fits the hypothesis.
3. Posterior Probability: This is the updated probability of the event occurring after considering the new data. It’s calculated using Bayes’ theorem.

Bayes’ Theorem:

Posterior Probability = (Likelihood * Prior Probability) / Evidence

Evidence is the total probability of observing the data, regardless of the hypothesis.

Example:

Let’s say, you’re trying to determine if a coin is fair. Your prior belief might be that it’s fair, so the probability of it landing heads or tails is 0.5.

1. Flip the coin 10 times and get 8 heads. This is your data.
2. Calculate the likelihood: The likelihood of getting 8 heads if the coin is fair is calculated using the binomial probability formula.
3. Update the posterior probability: Using Bayes’ theorem, you can calculate the updated probability of the coin being fair given the data.

If the posterior probability is significantly lower than the prior probability, it suggests that the coin might not be fair.

But why do we need this?

But isn’t it natural that we upgrade ourselves when we get new data? then, why do we need that theorem?

While it’s true that humans naturally update their beliefs based on new information, the Bayesian theorem provides a formal and systematic way to do this. Here’s why it’s valuable:

1. Quantifying Uncertainty: The Bayesian theorem allows us to quantify the degree of uncertainty associated with our beliefs. This is crucial in many fields, where precise measurements and predictions are essential.
2. Avoiding Cognitive Biases: Humans can be prone to cognitive biases that can distort our judgment. The Bayesian theorem can help us make more objective and rational decisions.
3. Complex Problems: For complex problems with many variables and uncertainties, the Bayesian theorem can provide a structured approach to reasoning and inference.
4. Automation: Bayesian methods can be automated, allowing computers to make decisions and predictions based on data.
5. Foundation for Machine Learning: Bayesian statistics is a fundamental tool in machine learning, enabling algorithms to learn from data and improve their performance over time.

Bayesian thinking is a powerful tool for making sense of a complex and ever-changing world. It’s used in fields like science, finance, medicine, and artificial intelligence to make better decisions and predictions.