What is Likelihood in Bayes' Theorem?

S7-SA3-0343

Grade Level:

Class 12

AI/ML, Physics, Biotechnology, FinTech, EVs, Space Technology, Climate Science, Blockchain, Medicine, Engineering, Law, Economics

Definition

What is it?

In Bayes' Theorem, 'Likelihood' tells us how probable our observed data is, assuming a particular hypothesis is true. It measures how well a specific hypothesis explains the evidence we have gathered.

Simple Example

Quick Example

Imagine you predict it will rain today (your hypothesis). If you see dark clouds and feel a cool breeze (your observed data), the likelihood is high because these signs often appear when it rains. If the sky is clear and sunny, the likelihood of your rain hypothesis being true is low.

Worked Example

Step-by-Step

Let's say a school's annual sports day has a 20% chance of being cancelled due to rain (P(Rain) = 0.2). We know that on rainy days, the ground is wet 90% of the time. On non-rainy days, the ground is wet only 10% of the time (due to sprinklers, etc.).

Now, imagine you wake up and see the ground is wet. We want to find the 'Likelihood' of rain given the wet ground, which is P(Wet Ground | Rain).

Step 1: Identify the event (hypothesis) we are assuming to be true. Here, it's 'Rain'.
---Step 2: Identify the evidence (observed data) we have. Here, it's 'Wet Ground'.
---Step 3: Look for the probability of the evidence happening, GIVEN that our assumed event is true. From the problem, we are given that on rainy days, the ground is wet 90% of the time.
---Step 4: So, P(Wet Ground | Rain) = 0.90.

Answer: The Likelihood of rain, given that the ground is wet, is 0.90 or 90%. This means if it rains, there's a 90% chance the ground will be wet.

Why It Matters

Likelihood helps AI systems understand patterns, like predicting if a customer will buy a product or if a medical test result indicates a disease. Engineers use it to build safer self-driving cars, and climate scientists use it to model future weather events, helping us prepare better for natural challenges.

Common Mistakes

MISTAKE: Confusing Likelihood P(Evidence | Hypothesis) with Prior Probability P(Hypothesis). | CORRECTION: Likelihood is about how likely the EVIDENCE is, given a hypothesis. Prior Probability is just how likely the hypothesis is BEFORE seeing any evidence.

MISTAKE: Thinking Likelihood is the same as Posterior Probability P(Hypothesis | Evidence). | CORRECTION: Likelihood is only one part of Bayes' Theorem. Posterior Probability is the FINAL probability of the hypothesis AFTER considering the evidence, using the full Bayes' formula.

MISTAKE: Assuming a high likelihood means the hypothesis is definitely true. | CORRECTION: A high likelihood means the evidence strongly supports the hypothesis, but it doesn't confirm it. Other factors (like the prior probability) also play a role.

Practice Questions

Try It Yourself

QUESTION: A new smartphone app has a bug that crashes 80% of the time when running on older phones. It crashes only 10% of the time on newer phones. If you assume you have an older phone, what is the Likelihood that the app will crash? | ANSWER: P(App Crash | Older Phone) = 0.80 or 80%

QUESTION: In a factory, 5% of the items produced are defective. A quality check machine correctly identifies a defective item 95% of the time. However, it also incorrectly flags a good item as defective 2% of the time. What is the Likelihood that the machine flags an item as defective, given that the item is actually defective? | ANSWER: P(Machine Flags Defective | Item is Defective) = 0.95 or 95%

QUESTION: A doctor is trying to diagnose a rare disease. The disease affects 1 in 1000 people. A diagnostic test for this disease gives a positive result 99% of the time if the person has the disease. It also gives a positive result 5% of the time if the person does NOT have the disease. If the doctor assumes a person HAS the disease, what is the Likelihood of getting a positive test result? What if the doctor assumes the person DOES NOT have the disease? | ANSWER: P(Positive Test | Has Disease) = 0.99 or 99%. P(Positive Test | Does NOT Have Disease) = 0.05 or 5%.

MCQ

Quick Quiz

Which of the following best describes 'Likelihood' in Bayes' Theorem?

The probability of a hypothesis being true before seeing any evidence.

The probability of observing the evidence, given that a specific hypothesis is true.

The final probability of a hypothesis being true after considering all evidence.

The overall probability of the evidence occurring.

The Correct Answer Is:

Likelihood (P(Evidence | Hypothesis)) measures how well the evidence fits a particular hypothesis. Option A is Prior Probability, Option C is Posterior Probability, and Option D is the Marginal Likelihood of Evidence.

Real World Connection

In the Real World

Imagine you're ordering food on an app like Swiggy or Zomato. The app uses likelihood to predict your estimated delivery time. If traffic is usually heavy at 7 PM (evidence) and it's 7 PM now (hypothesis), the likelihood of a longer delivery time is high. Similarly, online fraud detection systems use likelihood to identify suspicious transactions based on your past spending patterns.

Key Vocabulary

Key Terms

BAYES' THEOREM: A mathematical formula used to update the probability of a hypothesis as more evidence or information becomes available. | HYPOTHESIS: An idea or explanation for something that is tested with evidence. | EVIDENCE: The observed data or information collected. | PRIOR PROBABILITY: The initial probability of a hypothesis before any evidence is considered. | POSTERIOR PROBABILITY: The updated probability of a hypothesis after considering the evidence.

What's Next

What to Learn Next

Next, you should learn about 'Prior Probability' and 'Posterior Probability' in Bayes' Theorem. Understanding these will help you see how Likelihood combines with other factors to make powerful predictions and decisions, just like a real data scientist!