Artificial intelligence-based traffic flow prediction: a comprehensive review

Sayed, Sayed A.; Abdel-Hamid, Yasser; Hefny, Hesham Ahmed · 2023 · DOAJ

DOI: 10.1186/s43067-023-00081-6

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper presents a comprehensive review of artificial intelligence (AI) techniques applied to traffic flow prediction, a critical component of Intelligent Transportation Systems (ITS). The research is motivated by the growing urban congestion, rising road traffic fatalities, and the need for efficient mobility management in smart cities. Accurate traffic forecasting is essential for route planning, infrastructure design, and real-time traffic control. The authors aim to survey the most popular machine learning (ML) and deep learning (DL) methods used in this domain and identify the inherent obstacles and future directions for these technologies. The study categorizes AI approaches into machine learning and deep learning, further subdividing ML into supervised, unsupervised, and reinforcement learning. The review details specific algorithms within each category. Supervised learning methods discussed include Support Vector Machines (SVM), K-Nearest Neighbors (KNN), Logistic Regression, Linear Regression, Decision Trees (DT), Random Forests (RF), and Naïve Bayes. The authors explain the mechanisms, advantages, and limitations of each; for instance, SVMs are noted for their effectiveness in high-dimensional spaces but struggle with non-linear data without kernel tricks, while RFs mitigate the overfitting issues of single Decision Trees through ensemble voting. Unsupervised learning techniques covered include K-Means Clustering, Principal Component Analysis (PCA), and Latent Dirichlet Allocation (LDA), which are primarily used for dimensionality reduction and feature extraction. Reinforcement learning methods, such as Q-Learning and Monte Carlo Tree Search (MCTS), are described as goal-oriented approaches that learn through environmental interaction and reward maximization. The paper also contextualizes these modern AI methods against historical parametric models like Auto-Regressive Integrated Moving Average (ARIMA) and Kalman Filtering, noting that while these traditional methods rely on linear assumptions, traffic flow is inherently stochastic and non-linear. Consequently, nonparametric ML models and deep learning architectures—such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and Long Short-Term Memory (LSTM)—have emerged to capture complex temporal and spatial patterns in big data. The review highlights that deep learning allows for the extraction of complex properties from raw input through multiple layers, offering superior performance in handling the non-linear nature of traffic data compared to shallow neural networks or traditional statistical models. The significance of this work lies in its systematic classification of AI techniques for traffic prediction, providing a theoretical background for researchers and practitioners. By outlining the specific benefits and drawbacks of various algorithms—such as the computational cost of KNN or the interpretability of Decision Trees—the paper aids in selecting appropriate models for specific ITS applications. It underscores the shift from traditional parametric forecasting to data-driven AI solutions, emphasizing that the efficacy of ITS depends heavily on the quality of traffic data and the suitability of the prediction algorithm. The review serves as a foundational resource for understanding the current state of AI in transportation and identifies challenges that must be addressed to improve prediction accuracy and system reliability.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• anticipation