An Effective Algorithm for Association Rules Mining from Temporal Quantitative Databases

Background/Objectives: The objective of this paper is to find the relationships among the time of events in temporal quantitative databases. Methods/Statistical Analysis: The fuzzy sets are applied to both quantitative attributes of database and time distance among events. Then, Apriori algorithm with improvement is used to find out all frequent sequences. Final, the rules are pulled out from the sequences. Findings: The finding rules from the proposed algorithm help to predict quantity and occur time of items from other previous items. A rule can be a form "If large number of computers is sold, small number of modems will be bought in a medium period". In this paper, experiments are made to show performance of the proposed algorithm. Applications/Improvements: The findings are useful to forcast in market basket analysis, stock market analysis.