FUZZY WEIGHTED FP-GROWTH ALGORITHM WITH LINGUISTIC TERM THRESHOLDS

Efficiently discovering associations in transaction databases could be of enormous value to planning marketing or designing store layouts. However, most recent studies assumed that all items having the same weights with numerical values of the thresholds may not be considered of importance. In order to distinguish the significance of each item, a fuzzy weighted FP-growth (FWFP-growth) algorithm is proposed for generating fuzzy weighted association rules from transaction database, not to mention the replacement of linguistic terms for minimum support and minimum confidence are also more understandable for decision makers. As demonstrated with realistic database and different thresholds, the experimental results illustrate the proposed method is more efficient.