Automotive recall probability forecast based on T-S fuzzy neural network evaluation model

The recalls of defective automotive are one of the effective ways of removing automotive defects and improving the safety of transportation. Using the American NHTSA large data which include complaints data, defects data and recall data, there will be analyzed the amount of complaints and the attributions of defects by severity are influenced factors of recall. Next the recall will be forecast by the information of complaints, crash, fire, injured and fatality. It will be opened by the quantitative forecast for recall of automotive, and it is a potent means for the government´s decision of car recalls.