AY1516 T2 Team13 Natasha Studio Findings RuleMining
| EXPLORATORY DATA ANALYSIS | OTHER ANALYSIS | DATABASE CREATION | ASSOCIATION RULE MINING |
|---|
Member and Price Packages
From the results, 11 association rules were found. The rule with the highest confidence is also seen to be “04 Weeks (Full-Course)” to “06 Weeks (Full-Course)”. This does seem intuitive as members would be more likely to start trying shorter courses before moving on to longer courses.
One interesting thing to note is that only 4 association rules has a lift value higher than 1. Thus, this implies that the rule of “04 Open Classes” to “08 Open Classes” and vice versa has a positive correlation. For customers that have bought “04 Open Classes”, they are 1.23 times more likely to buy “08 Open Classes” and vice versa. Similarly, for “06 Weeks (Full-Course)” to “04 Weeks (Full-Course)’ and vice versa, there is a slightly positive correlation. This means that if “06 Weeks (Full-Course)” is bought, they are 1.03 times more likely to buy “04 Weeks (Full-Course). Overall, these results show that the member seem to be more likely to stay within the same type of price packages i.e. either within courses or open classes.
Next Steps
The next step would be to further calibrate the model to adjust maximum items and other parameters. Also, at this point, time has not been taken into account. Thus, our team is looking towards performing sequence discovery in SAS EM using “date purchased” as well. We would also be analysing using other target variables like “Package” (Genre) and “Course / Open & Level” with the MemberIDs for further analysis.
