engleski

Comparison of decision tree methods in decision making under risks and data mining

There are two methods called a decision tree. One belongs to the field of decision making in uncertainty and risks, and the other belongs to the field of data mining. Researchers and practitioners often mix those methods. In the decision making under risks, the decision-maker must decide on a certain issue at a certain moment, but the consequences of the decision will be visible in the future. Here, a decision tree is used at the level of visualization of the problem timeline but also to determine the decision that must be made. The data mining field deals with the discovery of new, valuable knowledge about the problem domain. A decision tree is one of the methods used for prediction and classification. A data analyst collects data about past and develops a predictive model to predict future events on new, unseen data. In this paper, we are presenting both decision tree methods through their elements, procedures, and demonstrations on examples. Finally, in comparison, we analyze the differences between those two methods and give guidelines to: (i) when to apply one or other approach, (ii) how to combine different approaches when problem domain requires combination.

Decision tree ; decision making ; risks ; data mining

nije evidentirano