Real Time Traffic Incident Detection by Using Twitter Stream Analysis
| Autor: | Umer Farooq, Zonia Fayyaz, Sidra Perveen, M. Khalid Ashraf, Nazifa Nazir, Maryam Afzaal, Khadija Akbar |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
050210 logistics & transportation
Focus (computing) Social network Point (typography) business.industry Event (computing) Computer science 05 social sciences computer.software_genre Support vector machine 03 medical and health sciences 0302 clinical medicine Ranking 0502 economics and business The Internet Social media 030212 general & internal medicine Data mining business computer |
| Zdroj: | Human Systems Engineering and Design ISBN: 9783030020521 IHSED |
| Popis: | Internet sites are sources of information for the detection of events, a special mention of traffic activity and accidental accidents or earthquake detection system. Because of the rapid growth of the last 20 years, there have been frequent traffic congestions in cities around the world. The increase in vehicles has caused a greater number of traffic events and, as a result, there are no common resources. We present a methodology for the acquisition, processing and classification of public Tweets with Natural Language Processing (NLP) techniques using the Vector Machine Support (SVM) algorithm, using text classification using social network data to detect incidents. Our view can detect tweets related to traffic, with an accuracy of 88.27%. In this document, we focus on a real-time monitoring system to detect traffic, for Twitter streams analysis by ranking of Twitter posts. We cannot even distinguish if an outdoor event throws traffic or not, multiplying the classification problem and correcting it by point 88.89%. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|