Abstract：The applicability of data mining on crashworthiness data of the occupant restraint system and the finite element data has been demonstrated in this paper. According to the characteristics of the CAE data, a framework for data preparation has been developed based on object oriented ideology. Decision tree and attribute selection arithmetic have been used to perform the data mining on the crash simulation data of 98 instances. The result of the data mining has given us a deeper insight into the interrelations between the key variations of the occupant restrain system during its design and the performance in crash simulation. Finally, the learned rules have been tested on the real data sets. The result of the testing showed that the learned rules is proper, and can been used for the design of the occupant restrain system.
2010, Vol. 29 
