AIM in Unsupervised Data Mining
|Publication Type||Book Section|
|Title||AIM in Unsupervised Data Mining|
|Abstract||This chapter explores the differences between association rules extracted
using the likelihood mining criterion (LMC) and rules extracted by using frequent item-set rule mining (FRM). LMC provides a change in perspective for rule selection, from a measure of frequency in the dataset, to a measure of relationship between the rule items. For illustration, this chapter presents the evaluation of qualitative differences between LMC and FRM rules with three examples: (1) a Basic rule mining scenario to illustrate LMC properties, (2) an analysis relating socio-economic information and chemical exposure data, and (3) mining behaviour routines in patients undergoing neurological rehabilitation. Results show that LMC is capable of extracting rare rules and does not suffer from support ilution. Furthermore, LMC focuses on the individual event generating processes, while FRM focuses on their commonalities.
|Book Title||Artificial Intelligence in Medicine|