MHB How can we find probability space and events

mathmari
Gold Member
MHB
Messages
4,984
Reaction score
7
Hey! :giggle:

Give a probability space and events $A$, $B$ and $C$ such that $P [C \mid A]>P [C \mid A \cup B]$.So we want that $$\frac{P[C\cap A]}{P[A]}>\frac{P[C\cap (A\cup B)]}{P[A\cup B]}=\frac{P[(C\cap A)\cup (C\cap B)]}{P[A\cup B]}$$ But how can we find these events ? :unsure:
 
Physics news on Phys.org
mathmari said:
Give a probability space and events $A$, $B$ and $C$ such that $P [C \mid A]>P [C \mid A \cup B]$.

So we want that $$\frac{P[C\cap A]}{P[A]}>\frac{P[C\cap (A\cup B)]}{P[A\cup B]}=\frac{P[(C\cap A)\cup (C\cap B)]}{P[A\cup B]}$$ But how can we find these events ?

Hey mathmari!

It looks as if we're supposed to find an example where the events could be anything.
For instance $A$ and $B$ could be genetic modifications, and $C$ could be whether a certain condition occurs.
Either way, we would need to assign example probabilities to the events and their intersections such that the given expression holds. 🤔
 
Klaas van Aarsen said:
It looks as if we're supposed to find an example where the events could be anything.
For instance $A$ and $B$ could be genetic modifications, and $C$ could be whether a certain condition occurs.
Either way, we would need to assign example probabilities to the events and their intersections such that the given expression holds. 🤔

So we consider $A$ to be the event that the genetic modification A occurs and $B$ to be the event that the genetic modification B occurs.
Let $P(A)=0.30$, $P(B)=0.40$, $P(A\mid C)=0.25$, $P(B\mid C)=0.35$, $P(C)=0.70$ $P(A\cap B)=0$ and so $P(A\cup B)=P(A)+P(B)=0.30+0.40=0.70$.

Then we have $P(C\cap A)=P(A\mid C)\cdot P(C)=0.25\cdot 0.70=0.175$, $P(C\cap B)=P(B\mid C)\cdot P(C)=0.35\cdot 0.70=0.245$ and $P[(C\cap A)\cup (C\cap B)]=P[C\cap A]+P[C\cap B]=0.175+0.245=0.42$ and so we get at the left side $$\frac{P[C\cap A]}{P[A]}=\frac{0.175}{0.30}\approx 0.583$$ and at the right side $$ \frac{P[C\cap (A\cup B)]}{P[A\cup B]}=\frac{P[(C\cap A)\cup (C\cap B)]}{P[A\cup B]} =\frac{0.42}{0.70} =0.6$$ If all the calculations are correct then the probabilities that I tokk are wrong... :unsure:
 
mathmari said:
So we consider $A$ to be the event that the genetic modification A occurs and $B$ to be the event that the genetic modification B occurs.
Let $P(A)=0.30$, $P(B)=0.40$, $P(A\mid C)=0.25$, $P(B\mid C)=0.35$, $P(C)=0.70$ $P(A\cap B)=0$ and so $P(A\cup B)=P(A)+P(B)=0.30+0.40=0.70$.

Then we have $P(C\cap A)=P(A\mid C)\cdot P(C)=0.25\cdot 0.70=0.175$, $P(C\cap B)=P(B\mid C)\cdot P(C)=0.35\cdot 0.70=0.245$ and $P[(C\cap A)\cup (C\cap B)]=P[C\cap A]+P[C\cap B]=0.175+0.245=0.42$ and so we get at the left side $$\frac{P[C\cap A]}{P[A]}=\frac{0.175}{0.30}\approx 0.583$$ and at the right side $$ \frac{P[C\cap (A\cup B)]}{P[A\cup B]}=\frac{P[(C\cap A)\cup (C\cap B)]}{P[A\cup B]} =\frac{0.42}{0.70} =0.6$$ If all the calculations are correct then the probabilities that I tokk are wrong...
Yep. It's the wrong way around, so we probably need to increase one probability and decrease another to match. 🤔

Btw, if everything would be correct, than the probability of each possible intersection must also be between $0$ and $1$, since otherwise we don't have a valid probability function. 🤔
 
When we consider the example https://mathhelpboards.com/threads/probability-chicken-legs.29067/ with $A=V_1$, $B=V_2$, $C=K$ then we get the following :

$P(V_1)=0.20$, $P(V_2)=0.10$, $P(K)=0.21$
$P(K\mid V_1)=0.80$
$P(K\mid V_1\cup V_2)$
$P(K\mid (V_1\cup V_2))=\frac{P(K\cap (V_1\cup V_2))}{P(V_1\cup V_2)}=\frac{P((K\cap V_1)\cup (K\cap V_2))}{P(V_1\cup V_2)}=\frac{P(K\cap V_1)+P(K\cap V_2)}{0.30}=\frac{0.16+0.04}{0.30}=\frac{0.20}{0.30}\approx 0.67$

So the inequality holds, right? :unsure:
 
mathmari said:
When we consider the example https://mathhelpboards.com/threads/probability-chicken-legs.29067/ with $A=V_1$, $B=V_2$, $C=K$ then we get the following :

$P(V_1)=0.20$, $P(V_2)=0.10$, $P(K)=0.21$
$P(K\mid V_1)=0.80$
$P(K\mid V_1\cup V_2)$
$P(K\mid (V_1\cup V_2))=\frac{P(K\cap (V_1\cup V_2))}{P(V_1\cup V_2)}=\frac{P((K\cap V_1)\cup (K\cap V_2))}{P(V_1\cup V_2)}=\frac{P(K\cap V_1)+P(K\cap V_2)}{0.30}=\frac{0.16+0.04}{0.30}=\frac{0.20}{0.30}\approx 0.67$

So the inequality holds, right?
Yep. Looks good to me. (Nod)
 
Klaas van Aarsen said:
Yep. Looks good to me. (Nod)

Great! Thank you! (Sun)
 
Back
Top