Abstract: A method for analyzing sets of temporal data using a computer is described. Each set of temporal data includes a plurality of records collected at a time unique to each such set and in which each record has a plurality of data items. The method includes the first step of creating data association rules for at least a plurality of sequential data sets wherein each association rule describes data items that are found together in some data records. An incidence proportion is then determined for each such association rule for each temporal data set and these incidence proportions are stored. The incidence proportions for each association rule are grouped into two aggregate proportions by a user defined window schedule that describes one or more pairs of a current window wc and a past window wp. The cumulative incidence proportions for each association rule for each past window wp are compared with the cumulative incidence proportion for the corresponding current window wc by a statistical test of two proportions.

Abstract: A method for analyzing sets of temporal data using a computer wherein each set of temporal data includes a plurality of records collected at a time unique to each such set and in which each record has a plurality of data items. The method includes the first step of creating data association rules for at least a plurality of sequential sets wherein each association rule represents data records having at least some common data items (100). A confidence factor is then determined for each such association rule and these confidence factors are stored in data partitions for the temporal data sets (102). The confidence factors for a selected data partition is then compared with the corresponding confidence factors of at least one other data partition (112), if available. When the confidence factor for the selected data partition varies from the corresponding confidence factor for the at least one other data partition exceeds a threshold value, an alert output signal is generated (114).