Assignment 2: Association Rule Mining

Start date 26 January, due 2 February beginning of class.

Exercises from the Book

Complete the following exercises from the book.

6.3
6.7
6.11
(From Han&Kamber Question 6.3(a):) For the following database,
1. Construct the FP-tree and mine rules using FP-growth having support ≥ 60% and confidence ≥ 80%:
  TID date items_bought
  
  T100 10/15/99 {K, A, D, B}
  
  T200 10/15/99 {D, A, C, E, B}
  
  T300 10/19/99 {C, A, B, E}
  
  T400 10/22/99 {B, A, D}
2. Suppose the following transactions are added to the database:
  TID date items_bought
  
  T500 10/27/99 {K, A, C}
  
  T600 10/28/99 {A, B, C}
  
  Can you figure out a more efficient means of updating the FP-tree and mining new rules than starting from scratch? You need not give a complete algorithm (although you may), but you should at least enumerate the difficulties/challenges (e.g., things you'd need to keep track of.)

TID	date	items_bought
T100	10/15/99	{K, A, D, B}
T200	10/15/99	{D, A, C, E, B}
T300	10/19/99	{C, A, B, E}
T400	10/22/99	{B, A, D}

TID	date	items_bought
T500	10/27/99	{K, A, C}
T600	10/28/99	{A, B, C}

Turning in assignment

Electronic submission preferred. Please email to clifton_nospam@cs_nojunk.purdue.edu . Pdf is the safest for capturing non-text (see ghostscript.) Hard copy is acceptable, please hand in at the beginning of class.