Practical Applications of Data Mining

2.5 Quantitative assoCiation rules 55

These two problems create the following conflicting situations: (1) large

intervals may generate rules that do not satisfy minimum confidence, and

(2) small intervals may generate rules that do not satisfy minimum support.

These situations can be resolved by considering all possible continuous ranges

over the values of the quantitative attribute or over the partitioned intervals

since the latter problem (minSup problem) disappears by combining adjacent

intervals or values. Although the minConf problem is still present, the infor-

mation loss is reduced by increasing the number of intervals that avoid the

minSup problem. This implies that the number of intervals must be increased

while simultaneously