Augmenting transportation-related recommendations through data mining

Abstract

This paper reports on the exploitation of data mining techniques during the formulation of purposeful association rules out of the transactions' database of a transportation management system. The rules' construction is performed through an elaborated version of the AprioriTid algorithm. The proposed algorithm is generic and capable to construct such rules by creating a large set of related items. The constructed rules can be used by the system's recommender module, which is responsible for providing recommendations to the associated users. The recommendation process takes into account the constructed rules and techniques that derive from the area of Collaborative Filtering (CF).