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

Manibardo, Eric L.; Laña, Ibai; Lobo, Jesús L.; Ser, Javier Del

doi:10.1109/ijcnn48605.2020.9207661

Cited by 5 publications

(3 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Both in the city and in highways, drivers tend to maintain a nominal speed whenever possible, so time series drops suddenly. Thereby, only the last timestamps provide information on this phenomena [165]. Results for New York and Seattle data sources corroborate this statement, where the performance degradation maintains a similarly decaying trend.…”

Section: B Results and Statistical Analysissupporting

confidence: 57%

“…Graph based, image based, together with some time series based model works, represent more than half of revised publications dealing with network-wide coverage solutions. While these studies usually concentrate on performing simultaneous predictions for multiple traffic network points, publications classified as point often put their effort on other specific issues like traffic signal processing [111], [132], the exploration of new data sources [59], [118], the improvement of performance under particular situations [103], [165] or missing data [47], [160].…”

Section: B Understanding Deep Learning Based Short-term Traffic Forec...mentioning

confidence: 99%

“…For this purpose, online learning techniques allow for the incremental update of the model when fed with new data, whereas concept drift handling approaches permit to adapt the behavior of the forecasting model to changing data distributions. Although the literature provides specific publications about this topic [165], [234], [235], [236], [237], it remains as a largely uncharted research area in traffic forecasting.…”

Section: A Actionability: Adaptive Models and Prediction Confidencementioning

confidence: 99%

See 2 more Smart Citations

Deep Learning for Road Traffic Forecasting: Does it Make a Difference?

Manibardo¹,

Laña²,

Ser³

2020

Preprint

Self Cite

View full text Add to dashboard Cite

Deep Learning methods have been proven to be flexible to model complex phenomena. This has also been the case of Intelligent Transportation Systems (ITS), in which several areas such as vehicular perception and traffic analysis have widely embraced Deep Learning as a core modeling technology. Particularly in short-term traffic forecasting, the capability of Deep Learning to deliver good results has generated a prevalent inertia towards using Deep Learning models, without examining in depth their benefits and downsides. This paper focuses on critically analyzing the state of the art in what refers to the use of Deep Learning for this particular ITS research area. To this end, we elaborate on the findings distilled from a review of publications from recent years, based on two taxonomic criteria. A posterior critical analysis is held to formulate questions and trigger a necessary debate about the issues of Deep Learning for traffic forecasting. The study is completed with a benchmark of diverse short-term traffic forecasting methods over traffic datasets of different nature, aimed to cover a wide spectrum of possible scenarios. Our experimentation reveals that Deep Learning could not be the best modeling technique for every case, which unveils some caveats unconsidered to date that should be addressed by the community in prospective studies. These insights reveal new challenges and research opportunities in road traffic forecasting, which are enumerated and discussed thoroughly, with the intention of inspiring and guiding future research efforts in this field.

show abstract

Section: B Results and Statistical Analysissupporting

confidence: 57%

Section: B Understanding Deep Learning Based Short-term Traffic Forec...mentioning

confidence: 99%

Section: A Actionability: Adaptive Models and Prediction Confidencementioning

confidence: 99%

See 1 more Smart Citation

Deep Learning for Road Traffic Forecasting: Does it Make a Difference?

Manibardo¹,

Laña²,

Ser³

2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

An active adaptation strategy for streaming time series classification based on elastic similarity measures

Oregi

Pérez

Ser

et al. 2022

Neural Comput & Applic

Self Cite

View full text Add to dashboard Cite

In streaming time series classification problems, the goal is to predict the label associated to the most recently received observations over the stream according to a set of categorized reference patterns. In online scenarios, data arise from non-stationary processes, which results in a succession of different patterns or events. This work presents an active adaptation strategy that allows time series classifiers to accommodate to the dynamics of streamed time series data. Specifically, our approach consists of a classifier that detects changes between events over streaming time series. For this purpose, the classifier uses features of the dynamic time warping measure computed between the streamed data and a set of reference patterns. When classifying a streaming series, the proposed pattern end detector analyzes such features to predict changes and adapt offline time series classifiers to newly arriving events. To evaluate the performance of the proposed scheme, we employ the pattern end detection model along with dynamic time warping based nearest neighbor classifiers over a benchmark of ten time series classification problems. The obtained results present exciting insights into the detection accuracy and latency performance of the proposed strategy.

show abstract