08.10.2016 Views

Foundations of Data Science

2dLYwbK

2dLYwbK

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Assume c < 1. Write x = I 1 + I 2 + · · · + I n where I i is the indicator variable indicating<br />

∑<br />

whether vertex i is an isolated vertex. Then E (x 2 ) = n E (Ii 2 ) + 2 ∑ E (I i I j ). Since I i<br />

i=1 i 1 there almost surely<br />

n<br />

are no isolated vertices, and when c < 1 there almost surely are isolated vertices.<br />

Hamilton circuits<br />

So far in establishing phase transitions in the G(n, p) model for an item such as the<br />

disappearance <strong>of</strong> isolated vertices, we introduced a random variable x that was the number<br />

<strong>of</strong> occurrences <strong>of</strong> the item. We then determined the probability p for which the expected<br />

value <strong>of</strong> x went from zero to infinity. For values <strong>of</strong> p for which E(x) → 0, we argued that<br />

with high probability, a graph generated at random had no occurrences <strong>of</strong> x. For values <strong>of</strong><br />

x for which E(x) → ∞, we used the second moment argument to conclude that with high<br />

probability, a graph generated at random had occurrences <strong>of</strong> x. That is, the occurrences<br />

that forced E(x) to infinity were not all concentrated on a vanishingly small fraction <strong>of</strong><br />

the graphs. One might raise the question for the G(n, p) graph model, do there exist<br />

items that are so concentrated on a small fraction <strong>of</strong> the graphs that the value <strong>of</strong> p where<br />

E(x) goes from zero to infinity is not the threshold? An example where this happens is<br />

Hamilton circuits.<br />

86

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!