Rules mining on hybrid electric vehicle consumer complaint database

Abstract

Abstract The hybrid electric vehicle (HEV) is a critical transportation disruptive technology that is expected to be widely adopted in the current and future marketplace. Many nations are promoting the success of HEVs. As the technologies and designs of these vehicles are significantly different from conventional vehicles, it is also important to understand the technical and body-related issues associated with these vehicles. This study used the National Highway Traffic Safety Administration’s vehicle owner’s complaint database to explore the potential issues associated with HEVs. The acquired dataset was divided into two groups based on their involvement in traffic crashes. The study applied association rule mining and text mining methods to analyze vehicle consumer complaint data. The results of association rule mining showed a significant association between hybrid electric all-wheel-drive vehicles manufactured between 2010 and 2021 that do not have anti-lock brakes and cruise control in the crash-related vehicle complaints dataset. Non-HEV vehicles, manufactured between 1992 and 1999, with cruise control and anti-braking systems as well as 5-10 cylinders, appeared frequently in the crash-related complaint dataset. Mileage-related issues and comparatively older HEVs (2000-2009) are dominant in non-crash-related data. The results from the text mining method show that brakes, mileage, failure, and crash are key features for consumer complaints related to crashes and brakes, battery, power, and recall are the key features for consumer complaints not related to crashes. The sentiment analysis results show slightly higher negative sentiments in complaint reports associated with crashes. The findings of this study can provide some insights into this unexplored research area.