빅데이터를 이용한 서울시 행복지수 분석 및 예측을 위한 실험 및 고찰

• 영문명: Forthcoming Big Data in Smart Cities: Experiment for Machine Learning Based Happiness Estimation in Seoul City
• 발행기관: 한국BIM학회
• 저자명: 신동윤(hin, Dongyoun) 송유미(Song, Yu-Mi)
• 간행물 정보: 『KIBIM Magazine』7권 1호, 1~8쪽, 전체 8쪽
• 주제분류: 공학 > 건축공학
• 파일형태: PDF
• 발행일자: 2017.03.30

영문 초록

Cities have complex system composed diverse activities. The activities in cities have complex relationship that creates diverse urban phenomena. Big Data is emerging technology in order to understand such complex network. This research aims to understand such relations by analysing the diverse city indexes. 28 indexes were collected in 25 of districts in Seoul city and analysed to find a weighted correlation. By defining the correlation values of certain years, it tries to predict the missed index values, “happiness” of each districts in other years. The result presents that the overall prediction accuracy 70.25%. However, for further discussion, the result is considered that this methods may not enough to use in practice, since the data has inconstant accuracy by different learning years.

목차

1. 연구의 배경 및 목적

2. 도시 데이터 수집

3. 가중치를 이용한 상관관계 분석

4. 학습결과를 이용한 예측

5. 가중치 값을 이용한 연도별 예측 결과

키워드

도시 빅데이터 기계학습 머신러닝 예측 모델 도시 지표 가중치 분석 Big data machine learning prediction model urban data model