A Multi Stream Feature Fusion Approach For Traffic Prediction

Authors

  • Dr.CHIRAPARAPU SRINIVASA RAO Author
  • SANABOYINA KEERTHI Author

Keywords:

decision-making, Support Vector Machines (SVM), Logistic Regression, Decision Tree, Classifier

Abstract

Traffic prediction plays a crucial role in 
urban planning, transportation 
management, and public safety. 
Traditional methods often rely on 
single data sources and simplistic 
models, leading to limited accuracy and 
reliability. This paper proposes a novel 
Multi-Stream Feature Fusion Approach 
for Traffic Prediction, leveraging 
diverse data streams and advanced 
machine learning algorithms. By 
integrating information from multiple 
sources such as historical traffic data, 
weather conditions, road infrastructure, 
and special events, the system enhances 
prediction accuracy and robustness. 
Feature fusion techniques, including 
concatenation and attention 
mechanisms, are employed to integrate
features from different streams 
effectively. Furthermore, various 
machine learning algorithms such as 
Support Vector Machines (SVM), 
Logistic Regression, and Decision Tree 
Classifier ,Naive bayes are utilized to 
train predictive models, capturing 
complex patterns in the data. Real-time 
prediction capabilities enable timely
decision-making, while visualization 
tools provide intuitive insights into 
predicted traffic patterns. Experimental 
results demonstrate the superiority of 
the proposed approach over traditional 
methods, highlighting its potential for 
improving traffic prediction accuracy 
and supporting efficient urban mobility 
management

Downloads

Published

11-06-2024

How to Cite

A Multi Stream Feature Fusion Approach For Traffic Prediction. (2024). International Journal of Engineering Research and Science & Technology, 20(2), 1125-1137. https://ijerst.org/index.php/ijerst/article/view/377