Inaugurated by IN-SPACe
ISRO Registered Space Tutor
S3-SA3-0447
What is a Posteriori Probability?
Grade Level:
Class 9
AI/ML, Data Science, Physics, Economics, Cryptography, Computer Science, Engineering
Definition
What is it?
A posteriori probability is the updated probability of an event happening AFTER we have observed some new information or evidence. It's like re-evaluating our chances based on fresh data, making our prediction more accurate. This updated probability is often calculated using Bayes' Theorem.
Simple Example
Quick Example
Imagine you initially think there's a 50% chance your favorite cricket team will win today's match. But then, you see the opposing team's star batsman got injured during warm-up. This new information (the injury) makes you update your belief that your team now has a higher chance of winning, say 70%. That 70% is an a posteriori probability.
Worked Example
Step-by-Step
Let's say a local bakery makes two types of sweets: Laddoo and Gulab Jamun. 60% of their sweets are Laddoos and 40% are Gulab Jamuns. We also know that 10% of Laddoos are slightly burnt, and 5% of Gulab Jamuns are slightly burnt. You pick a random sweet and find it is slightly burnt. What is the probability that it's a Laddoo?
1. **Identify Initial Probabilities (Prior):**
P(Laddoo) = 0.60 (60% are Laddoos)
P(Gulab Jamun) = 0.40 (40% are Gulab Jamuns)
2. **Identify Conditional Probabilities (Likelihood):**
P(Burnt | Laddoo) = 0.10 (10% of Laddoos are burnt)
P(Burnt | Gulab Jamun) = 0.05 (5% of Gulab Jamuns are burnt)
3. **Calculate Probability of the Evidence (Burnt Sweet):**
P(Burnt) = P(Burnt | Laddoo) * P(Laddoo) + P(Burnt | Gulab Jamun) * P(Gulab Jamun)
P(Burnt) = (0.10 * 0.60) + (0.05 * 0.40)
P(Burnt) = 0.06 + 0.02 = 0.08
4. **Apply Bayes' Theorem to find A Posteriori Probability:**
P(Laddoo | Burnt) = [P(Burnt | Laddoo) * P(Laddoo)] / P(Burnt)
5. **Substitute values:**
P(Laddoo | Burnt) = (0.10 * 0.60) / 0.08
P(Laddoo | Burnt) = 0.06 / 0.08
6. **Calculate the final probability:**
P(Laddoo | Burnt) = 0.75
**Answer:** The a posteriori probability that the burnt sweet is a Laddoo is 0.75 or 75%.
Why It Matters
A posteriori probability is super important in fields like AI/ML, where systems learn and update their understanding based on new data, for example, in facial recognition or spam detection. It helps scientists in Physics and Economics make better predictions and decisions by constantly refining their models with fresh observations. Engineers use it to improve system reliability and diagnose faults more accurately.
Common Mistakes
MISTAKE: Confusing a posteriori probability with prior probability. | CORRECTION: Prior probability is your initial belief BEFORE any new evidence. A posteriori probability is your updated belief AFTER considering new evidence.
MISTAKE: Not correctly identifying the 'evidence' or 'new information' in a problem. | CORRECTION: Always clearly pinpoint what new observation has occurred that makes you update your original probability.
MISTAKE: Mixing up P(A|B) with P(B|A) in calculations, especially when using Bayes' Theorem. | CORRECTION: Remember P(A|B) is the probability of A given B has occurred, which is different from P(B|A), the probability of B given A has occurred.
Practice Questions
Try It Yourself
QUESTION: A weather app initially predicts a 30% chance of rain today. In the morning, you see dark clouds. The app updates its prediction to 70%. Is the 70% an a priori or a posteriori probability? | ANSWER: A posteriori probability.
QUESTION: In a town, 2% of people have a rare disease. A test for this disease is 90% accurate (meaning it correctly identifies those with the disease 90% of the time, and correctly identifies those without the disease 90% of the time). If a person tests positive, what is the probability they actually have the disease? (Hint: Use Bayes' Theorem) | ANSWER: Approximately 15.66%.
QUESTION: A factory produces LED bulbs. 70% are from Machine A and 30% from Machine B. 5% of bulbs from Machine A are defective, and 10% of bulbs from Machine B are defective. If you pick a random bulb and find it is defective, what is the probability it came from Machine A? | ANSWER: P(Machine A | Defective) = [P(Defective | Machine A) * P(Machine A)] / [P(Defective | Machine A) * P(Machine A) + P(Defective | Machine B) * P(Machine B)] = (0.05 * 0.70) / (0.05 * 0.70 + 0.10 * 0.30) = 0.035 / (0.035 + 0.030) = 0.035 / 0.065 = 0.5385 or approximately 53.85%.
MCQ
Quick Quiz
Which of the following best describes a posteriori probability?
The initial probability of an event before any information.
The probability of an event after new evidence has been considered.
The probability of an event happening by chance.
The probability of two independent events occurring together.
The Correct Answer Is:
B
A posteriori probability is always about updating your belief or prediction after you've received new information or evidence. Option A describes prior probability.
Real World Connection
In the Real World
When you use a spam filter in your email, it constantly updates its understanding of what constitutes spam based on new emails it sees and your feedback. This learning process uses a posteriori probability. Similarly, recommendation systems on apps like YouTube or Amazon use it to suggest videos or products to you, updating their suggestions as you watch or buy more.
Key Vocabulary
Key Terms
PRIOR PROBABILITY: The initial probability of an event before any new information is observed. | EVIDENCE: New information or observation that influences the probability of an event. | BAYES' THEOREM: A mathematical formula used to calculate a posteriori probability. | CONDITIONAL PROBABILITY: The probability of an event occurring given that another event has already occurred.
What's Next
What to Learn Next
Now that you understand a posteriori probability, your next step should be to dive deeper into 'Bayes' Theorem'. This theorem is the mathematical backbone of calculating a posteriori probabilities and will help you solve more complex problems with confidence. Keep exploring!
