# Thread: Size of the intersection set

1. ## Size of the intersection set

Hello guys

I am thinking about a problem involving intersection of sets.

If I have for example a two sets A and B, and |A| is the number of elements in A and |B| the number of elements in B, then |A \cap B| is the number of elements in the intersection set A \cap B.

Now, if I add a third set C, with |C| elements, the number of elements in A \cap B \cap C will be naturally smaller (or equal) than in A \cap B. And if I'll add more sets, the number of elements in the intersection will keep decreasing.

Are there any researches / theorems about the rate of decrements ? I mean, if I take sets with numbers that are chosen by random, and I keep adding them, how strongly will the number of elements in the intersection converge to zero ?

Any insights / references will be most appreciated...

/cap means intersection, couldn't make latex work, don't know why...apologies...

2. ## Re: Size of the intersection set

Well it depends on what elements $C$ and $|A \cap B|$ have in common. You can't show that the intersection converges to zero, because each additional set might as well be equal to the intersection of A and B.

If A and B are fixed and are subsets of a universal set U, and if C, D, ... are subsets of fixed size, that's an entirely different question. Try small cases first.

3. ## Re: Size of the intersection set

What I am trying to prove is slightly different.

Let's say I have a set U which is a subset of the integers set Z. U contains around 10,000 elements.

Now I take a subset of U, let's call it A, and it has (just for illustration) 3000 elements. In the next step I will take another subset B, and it will have 2700 elements. I choose the elements of the subsets by random (!). Now I look at the intersection of A and B, it will contain (for example) 300 elements. In the next step I will take another random subset, C, and will look at the intersection of A, B and C. My intuition say that if I'll keep doing that, eventually the generalized intersection will converge to the empty set.

In the real world problem from which I took this challenge, I have seen the size of the generalized intersection decreases fast to 25 elements for 3 sets only. I do not have the 4th just yet, but I am trying to find a mathematical explanation to my intuition.

4. ## Re: Size of the intersection set

Ohh I see...one way to think of it, if any element n appears in the intersection of sets A,B,C, it must be contained in sets A,B,C. What is the probability of that happening?

I haven't thought about this for much time, but it looks like you'll probably get some sort of probability distribution based on the number and size of your subsets, and the size of the universal set U. Yes, intuitively, the intersection should "converge" to the empty set.