{"title":"EDSV: Emerging Defect Surveillance for Vehicles","authors":"Jiejun Xu, Daniel Xie, Tsai-Ching Lu, J. Cafeo","doi":"10.1145/3091478.3091521","DOIUrl":null,"url":null,"abstract":"We present early findings on building a proof of concept for an automated system to identify emerging trends regarding vehicle defects. The proposed system functions by continuously collecting and monitoring publicly available data from several heterogeneous channels ranging from online social media to vehicle enthusiast forums and consumer reporting sites. By mining the collected data, the system would provide real-time detection of ongoing consumer issues with vehicles. In addition, our system has special emphasis on detecting early signals prior to the widespread knowledge of the general public. One of the system components involves estimating a baseline statistical distribution governing the frequency of observing specific types of vehicle defective complaints from our data sources and subsequently identifying irregular deviations from this distribution. A web interface is made available to visualize descriptive statistics derived from various channels, with the intent to provide timely insights for human analysts.","PeriodicalId":165747,"journal":{"name":"Proceedings of the 2017 ACM on Web Science Conference","volume":"25 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2017-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 2017 ACM on Web Science Conference","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3091478.3091521","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
We present early findings on building a proof of concept for an automated system to identify emerging trends regarding vehicle defects. The proposed system functions by continuously collecting and monitoring publicly available data from several heterogeneous channels ranging from online social media to vehicle enthusiast forums and consumer reporting sites. By mining the collected data, the system would provide real-time detection of ongoing consumer issues with vehicles. In addition, our system has special emphasis on detecting early signals prior to the widespread knowledge of the general public. One of the system components involves estimating a baseline statistical distribution governing the frequency of observing specific types of vehicle defective complaints from our data sources and subsequently identifying irregular deviations from this distribution. A web interface is made available to visualize descriptive statistics derived from various channels, with the intent to provide timely insights for human analysts.