首页|Studies Conducted at University of Colorado on Machine Learning Recently Reporte d (Impacts of Increased Prediction Accuracy In Management Decisions: a Study of Full-depth Reclamation Pavements)
Studies Conducted at University of Colorado on Machine Learning Recently Reporte d (Impacts of Increased Prediction Accuracy In Management Decisions: a Study of Full-depth Reclamation Pavements)
扫码查看
点击上方二维码区域,可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Current study results on Machine Learn ing have been published. According to news reporting out of Boulder, Colorado, b y NewsRx editors, research stated, "Given the abundance of condition data regula rly collected for major roadways, machine learning has the potential to enhance pavement deterioration modeling. This is particularly important for recycling-ba sed rehabilitation techniques, such as full-depth reclamation (FDR), which lack accurate models of deterioration." Funders for this research include Colorado Department of Transportation, United States Department of Education Graduate Assistance in Areas of National Need Gra nt. Our news journalists obtained a quote from the research from the University of C olorado, "Previous studies have demonstrated the effectiveness of machine learni ng (ML) to predict pavement deterioration. However, the increased accuracy of th ese models often is reported using statistical metrics that pavement managers ca nnot easily relate to asset management decision-making. This paper quantifies th e impacts that increased accuracies in deterioration modeling have on relevant m etrics used in the management of pavement assets. The study analyzed the perform ance of full-depth-reclamation pavements and developed random forest models to e stimate roughness, rutting, and fatigue cracking. These random forest models wer e compared with mechanistic-empirical (M-E) models tuned to the same sites to qu antify differences in prediction accuracy, useful life, life-cycle costs, and lo ng-term performance. The tuned random forest deterioration models reduced errors by 90%-97% compared with the tuned M-E models. The r esults suggest that M-E predicts that FDR reaches the end of service life 8 year s sooner than do the random forest predictions. The long-term performance of FDR was found to be 28%-73% higher in a 10-year design l ife than M-E models predict."
BoulderColoradoUnited StatesNorth and Central AmericaCyborgsEmerging TechnologiesMachine LearningUniversit y of Colorado