Open AccessA Scheme for Predicting Energy Consumption in Smart Cities Using Machine Learning
Author(s) -
Noof Awad Alghamdi,
Israa Mohammed Budayr,
Samar Mohammed Aljehani,
Majed Aborokbah
Publication year - 2022
Publication title -
webology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.259
H-Index - 18
ISSN - 1735-188X
DOI - 10.14704/web/v19i1/web19230
Subject(s) - computer science , consumption (sociology) , energy consumption , urbanization , internet of things , aggregate (composite) , decision tree , outcome (game theory) , energy (signal processing) , transformation (genetics) , reading (process) , environmental economics , artificial intelligence , computer security , statistics , engineering , mathematics , economics , social science , materials science , mathematical economics , law , economic growth , chemistry , sociology , composite material , biochemistry , political science , gene , electrical engineering
Fluctuating result on weather condition throughout several decades became a global concern due to the direct or indirect effect on energy consumption, and that was well-defined in several sector. Research investigates the use of technology and the speed of obtaining information ، which helps in decision-making. This paper Emphasize the role of data science and their application to monitoring energy consumption, also, explain the importance used and challenges of Internet of Things (IoT). Thus, there is a global concern on data transformation from IoT devices when taking into account deferent weather variations. Cities are a critical part when of energy management, it presents the effect of urbanization and some of the success achievement in several cities around the world. Our Analysis indicate that three dissimilar types of sensors can detect massive amount of information up to four hundred thousand rows, compared to traditional methods for collecting data. The results depict the resilient of IOT performance which provide an aggregate of measures reach around 405,184 rows in a record time, with achieved accuracy up to 99% when implementing the decision tree algorithm, the outcome after applying the algorithm was vary 27.60 per-cent recorded by the first device while the other devices scored 26.14%,46.26% respectively, throughout different circumstances with continuous reading in a short period of times around 8 days.