Description

FpGrowth computes frequent itemsets given a set of transactions. The FP-Growth algorithm is described in Han et al., Mining frequent patterns without candidate generation.

Parameters

Name Description Type Required？ Default Value
itemsCol Column name of transaction items String ✓
minSupportCount Minimum support count Integer -1
minSupportPercent Minimum support percent Double 0.02
minConfidence Minimum confidence Double 0.05
maxPatternLength Maximum frequent pattern length Integer 10
maxConsequentLength Maximum consequent length Integer 1
minLift Minimum lift Double 1.0

Script Example

Code

  1. data = np.array([
  2.     ["A,B,C,D"],
  3.     ["B,C,E"],
  4.     ["A,B,C,E"],
  5.     ["B,D,E"],
  6.     ["A,B,C,D"],
  7. ])
  9. df_data = pd.DataFrame({
  10.     "items": data[:, 0],
  11. })
  13. data = dataframeToOperator(df_data, schemaStr='items string', op_type='batch')
  15. fpGrowth = FpGrowthBatchOp() \
  16.     .setItemsCol("items") \
  17.     .setMinSupportPercent(0.4) \
  18.     .setMinConfidence(0.6)
  20. fpGrowth.linkFrom(data)
  22. fpGrowth.print()
  23. fpGrowth.getSideOutput(0).print()

Results

Output

  1.     itemset  supportcount  itemcount
  2. 0         E             3          1
  3. 1       B,E             3          2
  4. 2       C,E             2          2
  5. 3     B,C,E             2          3
  6. 4         D             3          1
  7. 5       B,D             3          2
  8. 6       C,D             2          2
  9. 7     B,C,D             2          3
  10. 8       A,D             2          2
  11. 9     B,A,D             2          3
  12. 10    C,A,D             2          3
  13. 11  B,C,A,D             2          4
  14. 12        A             3          1
  15. 13      B,A             3          2
  16. 14      C,A             3          2
  17. 15    B,C,A             3          3
  18. 16        C             4          1
  19. 17      B,C             4          2
  20. 18        B             5          1

Output

  1.         rule  itemcount      lift  support_percent  confidence_percent  transaction_count
  2. 0       B=>A          2  1.000000              0.6            0.600000                  3
  3. 1       B=>D          2  1.000000              0.6            0.600000                  3
  4. 2       B=>C          2  1.000000              0.8            0.800000                  4
  5. 3       B=>E          2  1.000000              0.6            0.600000                  3
  6. 4       C=>A          2  1.250000              0.6            0.750000                  3
  7. 5       C=>B          2  1.000000              0.8            1.000000                  4
  8. 6     B,C=>A          3  1.250000              0.6            0.750000                  3
  9. 7       A=>C          2  1.250000              0.6            1.000000                  3
  10. 8       A=>B          2  1.000000              0.6            1.000000                  3
  11. 9       A=>D          2  1.111111              0.4            0.666667                  2
  12. 10    B,A=>D          3  1.111111              0.4            0.666667                  2
  13. 11    C,A=>B          3  1.000000              0.6            1.000000                  3
  14. 12    B,A=>C          3  1.250000              0.6            1.000000                  3
  15. 13    C,A=>D          3  1.111111              0.4            0.666667                  2
  16. 14  B,C,A=>D          4  1.111111              0.4            0.666667                  2
  17. 15      D=>A          2  1.111111              0.4            0.666667                  2
  18. 16      D=>B          2  1.000000              0.6            1.000000                  3
  19. 17    A,D=>B          3  1.000000              0.4            1.000000                  2
  20. 18    B,D=>A          3  1.111111              0.4            0.666667                  2
  21. 19    C,D=>B          3  1.000000              0.4            1.000000                  2
  22. 20    A,D=>C          3  1.250000              0.4            1.000000                  2
  23. 21    C,D=>A          3  1.666667              0.4            1.000000                  2
  24. 22  C,A,D=>B          4  1.000000              0.4            1.000000                  2
  25. 23  B,A,D=>C          4  1.250000              0.4            1.000000                  2
  26. 24  B,C,D=>A          4  1.666667              0.4            1.000000                  2
  27. 25      E=>B          2  1.000000              0.6            1.000000                  3
  28. 26    C,E=>B          3  1.000000              0.4            1.000000                  2