English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

This work focuses on classification over time series data. When a time seriesis generated by non-stationary phenomena, the pattern relating the series withthe class to be predicted may evolve over time (concept drift). Consequently,predictive models aimed to learn this pattern may become eventually obsolete,hence failing to sustain performance levels of practical use. To overcome thismodel degradation, online learning methods incrementally learn from new datasamples arriving over time, and accommodate eventual changes along the datastream by implementing assorted concept drift strategies. In this manuscript weelaborate on the suitability of online learning methods to predict the roadcongestion level based on traffic speed time series data. We draw interestinginsights on the performance degradation when the forecasting horizon isincreased. As opposed to what is done in most literature, we provide evidenceof the importance of assessing the distribution of classes over time beforedesigning and tuning the learning model. This previous exercise may give a hintof the predictability of the different congestion levels under target.Experimental results are discussed over real traffic speed data captured byinductive loops deployed over Seattle (USA). Several online learning methodsare analyzed, from traditional incremental learning algorithms to moreelaborated deep learning models. As shown by the reported results, whenincreasing the prediction horizon, the performance of all models degradeseverely due to the distribution of classes along time, which supports ourclaim about the importance of analyzing this distribution prior to the designof the model.

New Perspectives on the Use of Online Learning for Congestion Level Prediction over Traffic Data