Abstract: A temporal database which has time as the mandatory filed is considered to make the system more practical and realistic. Memory utilization is minimized by encoding the temporal database which involves several levels of data based on time. The values in each level are used in the encoded database to represent the transaction. This research proposes database encoding as a new presentation which can reduce the size of database and improve the efficiency of algorithms. An Encoded Temporal Database Method which identifies temporal association rules from an item set that consists of transactions with their corresponding valid time intervals. The application of Ant Colony Systems as a classification rule discovery is explored and probably to perform a flexible search over all possible logic combinations of the predicting attributes. The encoding database is used in a Temporal Data Mining System rather than classification rules in the sense of data mining. These results show to achieve both good predictive accuracy and reduce number of rules at the same time.
C. Balasubramanian, S. Saravanan, A. Shenbagarajan and K. Duraisamy, 2014. Encoded Temporal Database for ACS Classification Rule Discovery. International Journal of Soft Computing, 9: 338-347.