查看更多>>摘要：New research on Machine Learning is th e subject of a report. According to news originating from Camperdown, Australia, by NewsRx correspondents, research stated, "Understanding travellers' route pre ferences allows for the calculation of traffic flow on network segments and help s in assessing facility requirements, costs, and the impact of network modificat ions. Most research employs logit-based choice methods to model the route choice s of individuals, but machine learning models are gaining increasing interest." Our news journalists obtained a quote from the research from the University of S ydney, "However, all of these methods typically rely on a single ‘best' model fo r predictions, which may be sensitive to measurement errors in the training data . Moreover, predictions from discarded models might still provide insights into route choices. The ensemble approach combines outcomes from multiple models usin g various pattern recognition methods, assumptions, and/or data sets to deliver improved predictions. When configured correctly, ensemble models offer greater p rediction accuracy and account for uncertainties. To examine the advantages of e nsemble techniques, a data set from the I-35 W Bridge Collapse study in 2008, an d another from the 2011 Travel Behavior Inventory (TBI), both in Minneapolis-St. Paul (The Twin Cities) are used to train a set of route choice models and combi ne them with ensemble techniques. The analysis considered travellers' socio-demo graphics and trip attributes. The trained models are applied to two datasets, th e Longitudinal Employer-Household Dynamics (LEHD) commute trips and TBI morning peak trips, for validation. Predictions are also compared with the loop detector records on freeway links. Traditional Multinomial Logit and Path-Size Logit mod els, along with machine learning methods such as Decision Tree, Random Forest, E xtra Tree, AdaBoost, Support Vector Machine, and Neural Network, serve as the fo undation for this study. Ensemble rules are tested in both case studies, includi ng hard voting, soft voting, ranked choice voting, and stacking."
NETL
NSTL