Earticle

초록

영어

There are many obvious evidences supporting a correlation between school facility and student behavior and performance. With the increasing awareness of sustainable school facility, incorporation of various operation cost impact into the consideration of school facility management is attracting a lot of attention. So the Green-School Project in South Korea aims to transform existing deteriorated elementary, middle, and high school facilities into eco-friendly energy saving schools through environmentally friendly materials and techniques and full-scale renovation and repair work. However, the total number of educational facilities in South Korea as of 2015 is 11,590 (5,978 elementary schools, 3,219 middle schools, and 2,393 high schools). Overall reconstruction of these deteriorated educational facilities is realistically difficult. Expenditure by school systems must stay within the limit of their available funding. So in order to plan exact operating cost, this paper presents a prediction improving method of the amount of electricity consumption of elementary school in South Korea by using two regressions, i.e., SVR (Support Vector Regression) and GPR (Gaussian Process Regression) and outlier detection methods, EE (Elliptic Envelope) and EM (Expectation and Maximization) algorithms. As a result, this study enables school facility managers to straightforwardly predict the electricity consumption of elementary school. This method can also extend to prediction of the amount of electricity usage for middle school and high school as well as elementary school.

목차

Abstract
 1. Introduction
 2. Prediction Performance of the Amount of Electricity Consumption of Elementary School using Regression Algorithms
  2.1. Data Samples to Predict the Amount of Electricity Consumption of Elementary School
  2.2. Prediction Performance using SVR (Support Vector Machine) and GPR (Gaussian Process Regression)
 3. Elementary School Energy Prediction Performance Improvement by Applying Outlier Detection Method
  3.1. Outlier Detection Method
  3.2. Outlier Detection using EM Algorithm
  3.3. Scatter Plots of Predicted Value and True Value
 4. Conclusions
 Acknowledgments
 References

키워드

저자

• Hanguk Ryu [ Dept. of Architectural Engineering, Changwon National University ] Corresponding author
• Sebo Kim [ Dept. of Computer and Information Science and Engineering, University of Florida, Gainesville, FL, USA ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  International Journal of Smart Home

• 간기

  격월간

• pISSN

  1975-4094

• 수록기간

  2008~2016

• 십진분류

  KDC 505 DDC 605