0
$\begingroup$

I was reading about lift for frequent itemsets and came across this exercise here which was taken from a data mining textbook.

I tried plugging in the numbers, but could not get the 0.89 value for lift(Tea -> Coffee), can anyone point me to a direction of how they arrived at that number?

Improve this question
$\endgroup$
0
$\begingroup$

From all definitions that I can find about the lift (which are all equivalent because of Bayes), I always end up to a value of ~0.83. There is possibly an error in the page you cite.

For instance, from the data, my understanding is that the support of the rule tea $\rightarrow$ coffee is 0.15 (not 0.1). But this does not even explain the 0.89 value.

Improve this answer
$\endgroup$
  • $\begingroup$ I got ~0.83 as well by: P(tea and coffee) / (P (tea) * P(coffee) = 0.15 / (0.2 * 0.9) = 0.83. They might have got P(tea and coffee) to be 0.16 somehow, and that would give ~0.89. Anyway, your answer solidified mine, thank you! $\endgroup$ – Steven Than Oct 2 '18 at 16:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.