By: Bart Baesens, Seppe vanden Broucke
This QA first appeared in Data Science Briefings, the DataMiningApps newsletter as a “Free Tweet Consulting Experience” — where we answer a data science or analytics question of 140 characters maximum. Also want to submit your question? Just Tweet us @DataMiningApps. Want to remain anonymous? Then send us a direct message and we’ll keep all your details private. Subscribe now for free if you want to be the first to receive our articles and stay up to data on data science news, or follow us @DataMiningApps.
You asked: What is the lift value in association rule mining?
Our answer:
Consider the following example from a supermarket transactions database:
| Tea | Not tea | Total | |
| Coffee | 150 | 750 | 900 |
| Not coffee | 50 | 50 | 100 |
| Total | 200 | 800 | 1000 |
Let us now evaluate the association rule Tea => Coffee. The support of this rule is 100/1000 or 10%. The confidence of the rule is 150/200 or 75%. At first sight, this association rule seems very appealing given its high confidence. However, closer inspection reveals that the prior probability of buying coffee equals 900/1000 or 90%. Hence, a customer who buys tea is less likely to buy coffee than a customer about whom we have no information. The lift, also referred to as the interestingness measure, takes this into account by incorporating the prior probability of the rule consequent as follows:
A lift value less (larger) than 1 indicates a negative (positive) dependence or substitution (complementary) effect. In our example, the lift value equals 0.89, which clearly indicates the expected substitution effect between coffee and tea.
