Implementing Recommendation Algorithms for Decision Making Processes

This paper's contribution is placed into decision-making process research area. In our previous papers we showed how decision maker’s behavior can be captured in logs and how an aggregated decision data model (DDM) can be mined. We now introduce two recommendation algorithms that rely on a DDM. Each algorithm aims to steer the decision maker's actions towards a fully informed decision by suggesting the next action to be performed. The first algorithm is a Greedy approach that recommends the most frequent activities performed by other decision makers. The second algorithm is inspired from A* path finding algorithm. It finds a decision sub-objective and tries to guide the decision maker on the path to it. We eval-uate these algorithms by comparing them with each other and to the classical association rules approach. It is not our intention to recommend the better decision alternative. We want to make sure the decision makers made an informed decision by correctly and completely evaluating all alternatives. This is an extended version of the paper published at IE 2012 Conference.