Decision Rule Extraction for Maritime Accidents in Inland Rivers Based on Rough Set

In view of data mining for the decision-making of the inland waterway transportation, we study a rough set based approach to extract the decision rules with historical maritime accidents data in the inland rivers. In order to solve the NP-hard problem of attribute reduction in Rough Set Theory, the genetic algorithm based attribute reduction is proposed and described in detailed steps. Noting that there are generally multiple sorts of decision-making in various types of historical accidents, the decision problem thus first needs to be specified and its relevant attributes need to be determined to construct the decision table. Then, the relative minimal reduct of the decision table can be calculated using the proposed heuristic reduction algorithm, so that the decision rules can be obtained. The effectiveness of the proposed method is demonstrated with the decision rule extraction on collision and grounding accidents in the key segment of the Yangtze River.