THE ROLE OF DEEP LEARNING IN EXPLORING TRAFFIC PREDICTION TECHNIQUES DOI

Poonam Bhartiya,

Mukta Bhatele,

Akhilesh A. Waoo

et al.

ShodhKosh Journal of Visual and Performing Arts, Journal Year: 2024, Volume and Issue: 5(1)

Published: Jan. 31, 2024

This research paper delves into the pivotal role of deep learning in advancing traffic prediction techniques. With urban management becoming increasingly intricate, accurate short-term remains a cornerstone for effective congestion mitigation and transportation planning. Leveraging capabilities methodologies, this study systematically explores various models their applications predicting patterns. investigation clarifies advantages disadvantages approaches by looking at current developments, techniques, case examples. Moreover, it highlights avenues further development to enhance accuracy applicability learning-based systems, ultimately contributing evolution intelligent systems optimization mobility. Examine some most recent developments flow prediction. Convolutional neural networks (CNN), recurrent (RNNs), long (LONG-SNNNs), Stacked Auto Encoder (SAE), Restricted Boltzmann Machines (RBM), Term Memory (LSTM). These gradually extract higher-level information from raw input using numerous layers. Due complexity networks, created address challenge are examined. The reader is also informed on how aspects affect these which perform best specific circumstances.

Language: Английский

Advances in Traffic Congestion Prediction: An Overview of Emerging Techniques and Methods DOI Creative Commons
Aristeidis Mystakidis, Paraskevas Koukaras, Christos Tjortjis

et al.

Smart Cities, Journal Year: 2025, Volume and Issue: 8(1), P. 25 - 25

Published: Feb. 7, 2025

The ongoing increase in urban populations has resulted the enduring issue of traffic congestion, adversely affecting quality life, including commute duration, road safety, and local air quality. Consequently, recognizing forecasting underlying congestion patterns have become essential, with Traffic Congestion Prediction (TCP) emerging as an increasingly significant area study. Advancements Machine Learning (ML) Artificial Intelligence (AI), well improvements Internet Things (IoT) sensor technologies made TCP research crucial to development Intelligent Transportation Systems (ITSs). This review examines advanced TCP, emphasizing innovative methods their importance for ITS sector. paper provides overview statistical, ML, Deep (DL) approaches, ensembles that compose TCP. We examine several discuss relative absolute evaluation metrics from regression classification perspectives. Finally, we present overall step-by-step standard methodology is often utilized problems. By combining these elements, this highlights critical advancements challenges providing robust detailed information state-of-the-art solutions.

Language: Английский

Citations

1
Optimizing Traffic Accident Loss Predictions in China: Integrating Importance Indicator Screening with the Extra Trees Model for Greater Accuracy and Stability DOI
Rui Feng, Jian Liu,

Bin Lyu

et al.

Published: Jan. 1, 2025

Owing to the limited accuracy of traditional prediction methods, this work employs Extra Trees model from machine learning predict traffic accident losses through construction and data training. Utilizing importance-based indicator selection is substantially enhanced. The average for fatalities reaches 1.92%, while accuracies number accidents, property damage, injured persons exhibit varying degrees improvement, reaching 4.66%, 5.01%, 10.03%, respectively. This research identifies range multiple validations, ensuring stability. Under conditions stability, ana lysis calculation cumulative importance indicators during process are conducted based on theory, revealing stable ranking patterns across predictions. A comprehensive quantitative evaluation various in transportation sector performed, integrating relevance goodness fit into analysis. Results show that highway mileage, grade distance freight most closely associated with losses. study can serve as a reference assist formulating effective preventive measures.

Language: Английский

Citations

0
Optimizing Traffic Accident Loss Predictions in China: Integrating Importance Indicator Screening with the Extra Trees Model for Greater Accuracy and Stability DOI
Jian Liu, Rui Feng,

Bin Lyu

et al.

Published: Jan. 1, 2025

Owing to the limited accuracy of traditional prediction methods, this work employs Extra Trees model from machine learning predict traffic accident losses through construction and data training. Utilizing importance-based indicator selection is substantially enhanced. The average for fatalities reaches 1.92%, while accuracies number accidents, property damage, injured persons exhibit varying degrees improvement, reaching 4.66%, 5.01%, 10.03%, respectively. This research identifies range multiple validations, ensuring stability. Under conditions stability, ana lysis calculation cumulative importance indicators during process are conducted based on theory, revealing stable ranking patterns across predictions. A comprehensive quantitative evaluation various in transportation sector performed, integrating relevance goodness fit into analysis. Results show that highway mileage, grade distance freight most closely associated with losses. study can serve as a reference assist formulating effective preventive measures.

Language: Английский

Citations

0
Multilevel learning for enhanced traffic congestion prediction using anomaly detection and ensemble learning DOI

Mohammed A. Khasawneh,

Mustafa Daraghmeh, Anjali Awasthi

et al.

Cluster Computing, Journal Year: 2025, Volume and Issue: 28(3)

Published: Jan. 21, 2025

Language: Английский

Citations

0
A MACHINE LEARNING APPROACH FOR PREDICTIVE ANALYSIS OF TRAFFIC FLOW DOI

Poonam Bhartiya,

Mukta Bhatele,

Akhilesh A. Waoo

et al.

ShodhKosh Journal of Visual and Performing Arts, Journal Year: 2024, Volume and Issue: 5(5)

Published: May 31, 2024

Traffic congestion is a critical issue affecting urban areas globally, leading to significant economic and social costs. Predictive traffic flow analysis has emerged as promising solution mitigate enhance transportation efficiency. This paper proposes machine learning approach for predictive of flow, leveraging the wealth available data from various sources such sensors, GPS devices, cameras. paper's integrates historical with real-time information forecast future conditions accurately. employ combination techniques, including supervised unsupervised algorithms, model complex dynamics flow. Feature engineering techniques are applied extract meaningful features raw data, facilitating training models. Furthermore, it explores use advanced deep architectures, recurrent neural networks (RNNs) convolutional (CNNs), temporal spatial patterns. These models trained on large-scale datasets capture intricate relationships among different variables influencing Harnessing power can pave way smarter, more efficient systems that mobility reduce in environments.

Language: Английский

Citations

1
THE ROLE OF DEEP LEARNING IN EXPLORING TRAFFIC PREDICTION TECHNIQUES DOI

Poonam Bhartiya,

Mukta Bhatele,

Akhilesh A. Waoo

et al.

ShodhKosh Journal of Visual and Performing Arts, Journal Year: 2024, Volume and Issue: 5(1)

Published: Jan. 31, 2024

This research paper delves into the pivotal role of deep learning in advancing traffic prediction techniques. With urban management becoming increasingly intricate, accurate short-term remains a cornerstone for effective congestion mitigation and transportation planning. Leveraging capabilities methodologies, this study systematically explores various models their applications predicting patterns. investigation clarifies advantages disadvantages approaches by looking at current developments, techniques, case examples. Moreover, it highlights avenues further development to enhance accuracy applicability learning-based systems, ultimately contributing evolution intelligent systems optimization mobility. Examine some most recent developments flow prediction. Convolutional neural networks (CNN), recurrent (RNNs), long (LONG-SNNNs), Stacked Auto Encoder (SAE), Restricted Boltzmann Machines (RBM), Term Memory (LSTM). These gradually extract higher-level information from raw input using numerous layers. Due complexity networks, created address challenge are examined. The reader is also informed on how aspects affect these which perform best specific circumstances.

Language: Английский

Citations

0