id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
cord-025545-s6t9a7z8	Christantonis, Konstantinos	Using Classification for Traffic Prediction in Smart Cities	2020-05-06		.txt	text/plain	3291	199	56	This work focuses on analyzing different approaches regarding data manipulation in order to predict day-ahead traffic loads at random places around cities, based on weather conditions. Based on that, we used weather data collected from sensors installed around carefully chosen specific city spots for predicting the day-ahead traffic volume. To select the most appropriate locations to install sensors that either measure traffic loads or collect weather data, it is crucial to define their objective in advance. Our efforts focus on the question 'How can one exploit sensor data that are not personalized and create meaningful conclusions for the general public?' Deployment of smart city infrastructure requires a deep understanding of the traffic problem. Our approach, besides examining traffic predictability based on weather data, also aims to clarifying differences among locations. For example, if a sensor captures information every h(e.g. at 07:10, 08:10, 09:10 etc.), we computed and assigned the average value for each weather metric and the traffic load for that specific day period.	./cache/cord-025545-s6t9a7z8.txt	./txt/cord-025545-s6t9a7z8.txt