Session Ready
Exercise

Further refinement with lift

Once again, you report your results to the library: Use Twilight to promote Harry Potter, since the rule has a higher confidence metric. The library thanks you for the suggestion, but asks you to confirm that this is a meaningful relationship using another metric.

You recall that lift may be useful here. If lift is less than 1, this means that Harry Potter and Twilight are paired together less frequently than we would expect if the pairings occurred by random chance. As with the previous two exercises, the DataFrame books has been imported for you, along with numpy under the alias np.

Instructions
100 XP
  • Compute the support of {Potter, Twilight}.
  • Compute the support of {Potter}.
  • Compute the support of {Twilight}.
  • Compute the lift of {Potter} \(\rightarrow\) {Twilight}.