Bayes' Theorem Problem

62 Views Asked by At

I am having trouble trying to solve this problem. Anything helps!

Question: When doing a lab, the professor instructs students to get some rocks from a bucket and study them under a microscope. He says there are 90 sedimentary rocks (30 each of Chalk, Limestone, and Shale) and 180 Igneous rocks (60 each of Obsidian, Granite, and Diorite).

  1. You draw 3 rocks at random from the bucket. What is the probability that you got exactly 2 sedimentary rocks, given that you know you got at least one sedimentary rock?

  2. How many sedimentary rocks would you expect to receive in your 3-rock draw?

  3. Your friends tell you that Diorite, Obsidian, and Limestone have interesting and unique features when placed under a microscope. You now draw 18 rocks at random to check if your friends are right. What is the expected number of the interesting rocks in your 18-rock draw?

Attempted answer:

  1. I went into this using Bayes' Theorem.

$P(2S, 1I | 1S)$ = $\frac{P(1S | 2S, 1I)*P(2S, 1I)}{P(1S)}$

I know that P(1S) = 33.3% and P(2S, 1I) = $\frac{90}{270}$ * $\frac{89}{269}$ * $\frac{180}{268}$ ≈ 7.4%.

The issue that I have is P(1S | 2S, 1I). Is the answer to this 100% because you are always going to get 1 sedimentary rock given that you get 2 sedimentary rocks and 1 Igneous rock? Am I doing this right?

  1. This is an expected value question

E(1st draw) = $\frac{90}{270}$ * 1 = $\frac{1}{3}$

E(2nd draw) = $\frac{89}{269}$ * 1 = $\frac{89}{269}$

E(3rd draw) = $\frac{88}{268}$ * 1 = $\frac{88}{268}$

ΣE(draws) = 0.993

I feel like my numerator for the probabilities are wrong but I don't know how I should change that. How would I determine the expected values if the probabilities change?

  1. This is also an expected value question.

This question continues on the same concept as part 2, as the question uses the same fundamentals of part 2 where the probabilities would change given your draw. Would I approach in the same method or would I use a different approach?

1

There are 1 best solutions below

1
On BEST ANSWER

Am I doing this right?

No. The condition is that you have at least one sedimentary rock.

Let $S$ be the count for sedimentary rocks among the three you pick. This will have a hypergeometric distribution.

The probability for drawing exactly $2$ from $90$ sedimentary rocks and $1$ from $180$ igneous rocks when drawing $3$ from $270$ rocks is $$\mathsf P(S=2)=\left.\binom{90}2\binom{180}1\middle/\binom{270}3\right.$$

The probability for drawing at least $1$ from $90$ sedimentary rocks is one minus the probablity for the complementary event. $$\mathsf P(S\geq 1)=1-\mathsf P(S=0)$$

Put it together to find$$\mathsf P(S=2\mid S\geq 1) =\dfrac{\mathsf P(S=2)}{\mathsf P(S\geq 1)}$$


I feel like my numerator for the probabilities are wrong but I don't know how I should change that. How would I determine the expected values if the probabilities change?

The probabilities do not change. For example: take a standard deck of playing cards, select a card at random, now, what is the probability that it is an Ace?

Similarly, the probability that any particular rock among your three is sedimentary will be :

$$\mathsf E(S_1)=\mathsf E(S_2)=\mathsf E(S_3)=\dfrac{90}{270}$$


The third question is similar.