Week 9/5-9/12

Hi there!

So the first meeting on 9/5 was a major success. The group and I were given a few documents for literature review to get us started on understanding a general idea on what we were doing. We read:

  • Association Rules-Support and Confidence
  • Mining Association Rules
  • Data Mining-Apriori Algortihm

After reading these, we came back and talked about what we learned/were still confused on. While the concepts repeated each other in the documents, they provided different ways of explaining things like confidence, support, and lift along with the formulas to calculate all three.

Next we were assigned to recommend deals to “Alice”. This fictional character was interested in one deal, and using data from 4 other people, we worked on finding another deal that she may like.

  1. First, we worked on assigning confidence and support values to the deal transactions. For some of the deals, it was pretty straightforward. In order to calculate confidence, count the number of “interested” markers there were, use that as the denominator, then count how many of the second deal “interested” markers matched those of the initial deal. To calculate support, use the same numerator in the confidence and put it over the number of transactions. However, when deals had a “missing information” marker in the denominator, it made things tricky…this is illegal math! We still are unsure how to proceed with this…
  2. Second, I used the numerical values from before to eliminate deals that I knew she wouldn’t like (resulted in a 0 for the confidence and support). Then I put them in order from highest value to lowest. As a group, we were unsure what the question was asking, so no guarantee if it was the right method.
  3. Finally we selected the highest combination of confidence and support values in order to recommend our fictional shopper deals that we were sure she would like.

For next week we were asked to read up on lift, another comparison equation, and to see if that changes our answers at all.

If you’ve made it this far, thanks for reading! If not, I get it…it looks like a lot! Talk to you all next week!

–Alyssa