1 / 26

Association Analysis (3)

Association Analysis (3). FP-Tree/FP-Growth Algorithm. Use a compressed representation of the database using an FP-tree Once an FP-tree has been constructed, it uses a recursive divide-and-conquer approach to mine the frequent itemsets. Building the FP-Tree

micah
Download Presentation

Association Analysis (3)

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Association Analysis (3)

  2. FP-Tree/FP-Growth Algorithm • Use a compressed representation of the database using an FP-tree • Once an FP-tree has been constructed, it uses a recursive divide-and-conquer approach to mine the frequent itemsets. Building the FP-Tree • Scan data to determine the support count of each item. Infrequent items are discarded, while the frequent items are sorted in decreasing support counts. • Make a second pass over the data to construct the FP­tree. As the transactions are read, before being processed, their items are sorted according to the above order.

  3. First scan – determine frequent 1-itemsets, then build header

  4. null B:1 A:1 null B:2 C:1 A:1 D:1 FP-tree construction After reading TID=1: After reading TID=2:

  5. null B:8 A:2 A:5 C:3 C:1 D:1 C:3 D:1 D:1 E:1 D:1 E:1 D:1 E:1 FP-Tree Construction Transaction Database Header table Chain pointers help in quickly finding all the paths of the tree containing some given item.

  6. FP-Tree size • The size of an FP­tree is typically smaller than the size of the uncompressed data because many transactions often share a few items in common. • Best­case scenario: • All transactions have the same set of items, and the FP­tree contains only a single branch of nodes. • Worst­case scenario: • Every transaction has a unique set of items. • As none of the transactions have any items in common, the size of the FP­tree is effectively the same as the size of the original data. • The size of an FP­tree also depends on how the items are ordered. • If the ordering scheme in the preceding example is reversed, • i.e., from lowest to highest support item, the resulting FP­tree probably is denser (shown in next slide). • Not always though…ordering is just a heuristic.

  7. An FP­tree representation for the data set with a different item ordering scheme.

  8. FP-Growth (I) • FP­growth generates frequent itemsets from an FP­tree by exploring the tree in a bottom­up fashion. • Given the example tree, the algorithm looks for frequent itemsetsending in E first, followed by D, C, A, and finally, B. • Since every transaction is mapped onto a path in the FP­tree, we can derive the frequent itemsets ending with a particular item, say, E, by examining only the paths containing node E. • These paths can be accessed rapidly using the pointers associated with node E.

  9. null null B:3 B:8 A:2 A:2 A:5 C:3 C:3 C:1 C:1 D:1 D:1 C:3 D:1 D:1 E:1 E:1 D:1 D:1 E:1 E:1 D:1 E:1 E:1 Paths containing node E

  10. Conditional FP-Tree for E • We now need to build a conditional FP-Tree for E, which is the tree of itemsets ending in E. • It is not the tree obtained in previous slide as result of deleting nodes from the original tree. • Why? Because the order of the items change. • In this example, C has a higher count than B.

  11. Header table null The conditional FP-Tree for E B:3 A:2 C:3 C:1 D:1 null The new header C:3 C:1 A:1 E:1 D:1 E:1 B:3 E:1 A:1 D:1 The set of paths containing E. Insert each path (after truncating E) into a new tree. D:1 Conditional FP-Tree for E Adding up the counts for D we get 2, so {E,D} is frequent itemset. We continue recursively. Base of recursion: When the tree has a single path only.

  12. FP-Tree Another Example Transactions Transactions with items sorted based on frequencies, and ignoring the infrequent items. Freq. 1-Itemsets. Supp. Count 2 A B C E F O A C G E I A C D E G A C E G L E J A B C E F P A C D A C E G M A C E G N A C E B F A C G E A C E G D A C E G E A C E B F A C D A C E G A C E G

  13. FP-Tree after reading 1st transaction A C E B F A C G E A C E G D A C E G E A C E B F A C D A C E G A C E G Header null A:1 C:1 E:1 B:1 F:1

  14. FP-Tree after reading 2nd transaction A C E B F A C G E A C E G D A C E G E A C E B F A C D A C E G A C E G Header null A:2 C:2 G:1 E:1 B:1 F:1

  15. FP-Tree after reading 3rd transaction A C E B F A C G E A C E G D A C E G E A C E B F A C D A C E G A C E G Header null A:2 E:1 C:2 G:1 E:1 B:1 F:1

  16. FP-Tree after reading 4th transaction A C E B F A C G E A C E G D A C E G E A C E B F A C D A C E G A C E G Header null A:3 E:1 C:3 G:1 E:2 G:1 B:1 F:1 D:1

  17. FP-Tree after reading 5th transaction A C E B F A C G E A C E G D A C E G E A C E B F A C D A C E G A C E G Header null A:4 E:1 C:4 G:1 E:3 G:2 B:1 F:1 D:1

  18. FP-Tree after reading 6th transaction A C E B F A C G E A C E G D A C E G E A C E B F A C D A C E G A C E G Header null A:4 E:2 C:4 G:1 E:3 G:2 B:1 F:1 D:1

  19. FP-Tree after reading 7th transaction A C E B F A C G E A C E G D A C E G E A C E B F A C D A C E G A C E G Header null A:5 E:2 C:5 G:1 E:4 G:2 B:2 F:2 D:1

  20. FP-Tree after reading 8th transaction A C E B F A C G E A C E G D A C E G E A C E B F A C D A C E G A C E G Header null A:6 E:2 C:6 G:1 D:1 E:4 G:2 B:2 F:2 D:1

  21. FP-Tree after reading 9th transaction A C E B F A C G E A C E G D A C E G E A C E B F A C D A C E G A C E G Header null A:7 E:2 C:7 G:1 D:1 E:5 G:3 B:2 F:2 D:1

  22. FP-Tree after reading 10th transaction A C E B F A C G E A C E G D A C E G E A C E B F A C D A C E G A C E G Header null A:8 E:2 C:8 G:1 D:1 E:6 G:4 B:2 F:2 D:1

  23. Conditional FP-Trees Build the conditional FP-Tree for each of the items. For this: • Find the paths containing on focus item. With those paths we build the conditional FP-Tree for the item. • Read again the tree to determine the new counts of the items along those paths. Build a new header. • Insert the paths in the conditional FP-Tree according to the new order.

  24. Conditional FP-Tree for F null Header null New Header A:2 A:8 C:2 C:8 E:6 E:2 B:2 B:2 F:2 There is only a single path containing F

  25. Recursion • We continue recursively on the conditional FP-Tree for F. • However, when the tree is just a single path it is the base case for the recursion. • So, we just produce all the subsets of the items on this path merged with F. {F} {A,F} {C,F} {E,F} {B,F} {A,C,F}, …, {A,C,E,F} null New Header A:2 C:2 E:2 B:2

  26. null A:8 C:8 D:1 E:6 G:4 D:1 Paths containing D after updating the counts Conditional FP-Tree for D null New Header A:2 C:2 The other items are removed as infrequent. • The tree is just a single path; it is the base case for the recursion. • So, we just produce all the subsets of the items on this path merged with D. • {D} {A,D} {C,D} {A,C,D} • Exercise: Complete the example.

More Related