An Improved Deep Neural Network-Based Predictive Model for Traffic Accident’s Severity Prediction

Sarita Simaiya, Umesh Kumar Lilhore, Himanshu Pandey, Naresh Kumar Trivedi, Abhineet Anand, Jasminder Sandhu

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Scopus citations

Abstract

In recent decades, a vehicle crash has become a worldwide issue and labeled it the twelfth prominent reason for death worldwide. Artificial intelligence-based techniques, i.e., machine learning and deep learning play a vital role in various aspects of modern society and accidental predication is one of them. This research presents an accidental prediction model based on an improved deep neural network (IDNN). A DNN model mainly contains various hidden layers toward nodes. The proposed IDNN model includes two modules. The first model is based on an unsupervised feature learning interface to recognize operational networks and correlation factors. The second model is based on a supervised optimization subsystem with an extended negative binomial distribution which helps in forecasting road crashes. The proposed model was simulated over traffic datasets collected from the online data sources, and various performance measuring parameters have been calculated, i.e., precision, accidental detection. The experimental results demonstrate strengthen of the proposed model over the existing model.

Original languageEnglish
Title of host publicationAmbient Communications and Computer Systems - Proceedings of RACCCS 2021
EditorsYu-Chen Hu, Shailesh Tiwari, Munesh C. Trivedi, K. K. Mishra
PublisherSpringer Science and Business Media Deutschland GmbH
Pages181-190
Number of pages10
ISBN (Print)9789811679513
DOIs
StatePublished - 1 Jan 2022
Externally publishedYes
EventInternational Conference on Recent Advancement in Computer, Communication, and Computational Sciences, RACCCS 2021 - Ajmer, India
Duration: 20 Aug 202121 Aug 2021

Publication series

NameLecture Notes in Networks and Systems
Volume356
ISSN (Print)2367-3370
ISSN (Electronic)2367-3389

Conference

ConferenceInternational Conference on Recent Advancement in Computer, Communication, and Computational Sciences, RACCCS 2021
Country/TerritoryIndia
CityAjmer
Period20/08/2121/08/21

Keywords

  • AI
  • Accident prediction
  • Improved deep neural network extended binomial distribution
  • Road safety

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Signal Processing
  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'An Improved Deep Neural Network-Based Predictive Model for Traffic Accident’s Severity Prediction'. Together they form a unique fingerprint.

Cite this