View : 255 Download: 0

Full metadata record

DC Field Value Language
dc.contributor.advisor유재근-
dc.contributor.author박유연-
dc.creator박유연-
dc.date.accessioned2023-02-24T16:30:57Z-
dc.date.available2023-02-24T16:30:57Z-
dc.date.issued2023-
dc.identifier.otherOAK-000000201949-
dc.identifier.urihttps://dcollection.ewha.ac.kr/common/orgView/000000201949en_US
dc.identifier.urihttps://dspace.ewha.ac.kr/handle/2015.oak/264318-
dc.description.abstractClustering on distances of the functional time series data allows the dimension reduction and prediction of the infinite dimension data. This paper applies the specific technique, CATS (Clustering After Transformation and Smoothing) by Serban and Wasserman(2005) for the prediction of the monthly traffic volume. This paper conducts the transformation of time series data into the Fourier domain and data smoothing using 10 bases. Then it clusters the transformed functional data into 4 groups using the K means clustering method according to distances. Next, it extracts the data randomly from those 4 clusters before the prediction of the monthly traffic volume through its time and spatial information. In the prediction phase, it uses kriging to calculate the weights by the difference in distances between each point and the unobserved point. As a result, it predicts the estimated traffic volume.;시계열 함수형 데이터의 공간 정보, 즉 거리에 따른 군집화를 통해 차원축소 및 무한 차원의 데이터의 분류와 예측이 가능하다. 본 연구는 K-means 군집화 방법 기반 으로 함수형 데이터 분석을 통해 실제 월간 교통량을 예측한다. 교통량 시계열 데이터를 Fourier Spline을 통해 함수형 데이터로 변환한 뒤, 군집화 과정에서 각 데이터가 위치한 거리에 따라 네개의 군집으로 분류한다. 분류된 각 군집에서 임의로 데이터를 추출해, 추출된 데이터의 시간 및 공간적 정보를 토대로 월 교통량 예측을 수행한다.-
dc.description.tableofcontents1. Introduction 1 2. Methodology 3 2.1 CATS: Clustering After Transformation and Smoothing 3 2.1.1 Transformation 3 2.1.2 Smoothing 4 2.1.3 Clustering of Functional Data 5 2.2 Kriging 6 3. Data Analysis 8 3.1 Data Description 8 3.2 Application of Simple CATS 9 3.3 Functional Data Prediction 16 3.4 The Results 24 4. Discussion 26 References 27-
dc.formatapplication/pdf-
dc.format.extent3222546 bytes-
dc.languageeng-
dc.publisher이화여자대학교 대학원-
dc.subject.ddc500-
dc.titleSpatio-Temporal Functional Data Prediction Using CATS with Functional K-Means Clustering on the Traffic Volume Data-
dc.typeMaster's Thesis-
dc.creator.othernamePark, Yu Yeon-
dc.format.pageiii, 29 p.-
dc.identifier.thesisdegreeMaster-
dc.identifier.major대학원 통계학과-
dc.date.awarded2023. 2-
Appears in Collections:
일반대학원 > 통계학과 > Theses_Master
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

BROWSE