Bayesian formula for weather exercise

187 Views Asked by At

If it is nice weather on one day, the probability that it is going to be nice again the next is $13/15$.

If it is raining on one day, the prob. that it is going to be raining again the next day is $4/5$

Now $S_k$ is the event: "k-th day is nice" and we assume $\mathbb P(S_k)=\mathbb P(S_{k+1})\forall k\in\mathbb Z$ and $\mathbb P(S_{k+1}\cap S_{k-1}| S_k)=\mathbb P(S_{k+1}| S_k)\mathbb P(S_{k-1}| S_k)$

(a) What is $\mathbb P(S_{k-1}| S_k)$ ? I think we can use bayesian formula $\mathbb P(S_{k-1}| S_k)=\frac{\mathbb P(S_{k}| S_{k-1})\mathbb P(S_{k-1})}{\mathbb P(S_k)}$ and $\mathbb P(S_{k}| S_{k-1})=13/15$, but what about the other terms? Here the next point

(b) What is the probability that there is nice weather on a specific day?

(c) Probability that after a nice day we have two more nice days?

1

There are 1 best solutions below

0
On BEST ANSWER

Would perhaps $P(S_k)=P(S_{k+1})$ imply $P(S_{k-1})=P(S_k)$? If so, then $P(S_{k-1}|S_k)$ reduces to $P(S_k|S_{k-1})=13/15$. But that doesn't help with question b)...

In c) you have: $$ P(S_{k+2}\cap S_{k+1}|S_k)=P(S_{k+1}|S_k)P(S_{k+2}|S_k, S_{k+1})=P(S_{k+1}|S_k)P(S_{k+2}|S_{k+1})=[P(S_{k+1}|S_k)]^2=\left(\frac{13}{15}\right)^2 $$ which follows from your statement $$ P(S_{k+1}\cap S_{k-1}|S_k)=\left[P(S_{k-1}|S_k)P(S_{k-1}|S_k, S_{k+1})\right]=P(S_{k-1}|S_k)P(S_{k-1}|S_k) $$ where the middle part in brackets is the implied step. It tells us that, conditional on $S_k$, $S_{k-1}$ and $S_{k+1}$ are independent. I'm not 100 % sure, but maybe this can inspire you at least. Good luck!