A Nested Clustering Technique for Freeway Operating Condition Classification

This article introduces a nested clustering technique and its application to the analysis of freeway operating condition. A clustering model is developed using the traffic data (flow, speed, occupancy) collected by the detectors and aggregated to 5-minute increments. An optimum fit of the statistical characteristics of the data set is provided by the model based on the Bayesian Information Criterion and the ratio of changes in dispersion measurement. This technique is flexible in determining the number of clusters based on the statistical characteristics of the data. Tests on multiple sites with varying operating conditions have attested to its effectiveness as a data mining tool for the analysis of freeway operating condition.