Table 4.
Category | Subsample A | Subsample B | Subsample C |
---|---|---|---|
Number of subjects | 2,741 | 2,736 | 2,132 |
Number of chemicals | 29 | 37 | 40 |
Maximum weight | 476,883.0 | 426,061.1 | 413,068.1 |
Minimum weight | 14,002.7 | 13,975.1 | 12,659.3 |
Sum of weights | 258,281,689.4 | 272,911,664.0 | 226,021,580.6 |
Records needed | 18,445.1 | 19,528.5 | 17,854.1 |
Note: The number of subjects and the summary statistics for the subsample weights only reflect those subjects that met the criteria described in “Methods” section. That is, some NHANES 2009–2010 subjects were omitted from consideration because they did not meet age requirements for certain laboratory analyses or because chemical concentration information was incomplete. As discussed in the text, we preprocessed the raw data to obtain subject-chemical databases before conducting frequent itemset mining (FIM). Part of this process entailed duplicating subject records to reflect subsample weights. The final row in this table gives the total number of records needed (after duplication) so that each record corresponding to a subject with the minimum weight would occur exactly once in the transaction database.