I have a series of random variables, where the expected value is n4. I want to prove, with Chernoff bounds, that the probability that the actual value is less than (1−ϵ)n8 is very small.
I am unsure as to how to approach this problem. It is immediately obvious that n/8 is half of n/4. I don't believe solving for epsilon is the right approach for this, however. How do I approach this?
