Optimized Traffic Flow Prediction Using Sparrow Search Algorithm and LSTM Fine-Tuned with Hybrid Sandpiper Optimization

In order to implement intelligent transportation systems effectively, precise and up-to-date data on traffic flows is crucial. To have officially entered the age of big data in transportation, thanks to the explosion of traffic data in recent years. Unfortunately, many practical applications predict...

Full description

Saved in:
Bibliographic Details
Published in:2025 International Conference on Computational Robotics, Testing and Engineering Evaluation (ICCRTEE) pp. 1 - 6
Main Authors: Hafeez, Shaik Abdul, R, Mahaveerakannan
Format: Conference Proceeding
Language:English
Published: IEEE 28.05.2025
Subjects:
Accuracy
Adaptation models
Computational modeling
Feature extraction
Hybrid Sandpiper Optimization
Intelligent Transport System
Internet of Things
Long Short-Term Memory
Optimization methods
Prediction algorithms
Predictive models
Real-time systems
Sparrow Search Algorithm
Traffic flow prediction
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In order to implement intelligent transportation systems effectively, precise and up-to-date data on traffic flows is crucial. To have officially entered the age of big data in transportation, thanks to the explosion of traffic data in recent years. Unfortunately, many practical applications prediction, which primarily rely on shallow traffic prediction models. An essential part of ITS, traffic flow prediction allows for better traffic management, congestion reduction, and route optimization. In accuracy of the model, this study suggests a state-of-the-art predictive framework that combines SSA for feature selection with an LSTM network that has been fine-tuned with HSO. Through the elimination of redundant data, selection of the most relevant traffic-related factors, and improvement of computing performance, SSA successfully minimizes the feature space. Faster convergence and higher generalization are achieved by optimizing the LSTM model using HSO, which adaptively fine-tunes hyperparameters. The model is well-known for its aptitude to grasp temporal dependencies. The suggested SSA-HSO-LSTM model outperforms conventional feature selection and optimization methods, rendering it adept at managing unpredictable and non-linear traffic conditions. Results from experiments confirm the method's robustness, outperforming both traditional approaches and baseline deep learning models. The suggested method provides a trustworthy and scalable answer to the problem of real-time traffic forecasting, despite the fact that computational complexity is still an issue. Improving real-time traffic prediction and management through the use of the Internet of Things (IoT) besides edge computing, as well as incorporating real-time adaptive learning techniques, could be the subject of future research.
DOI:10.1109/ICCRTEE64519.2025.11053061