The Comparison of the two-stage EWMA approach with conventional EWMAs for forecasting time series data
The Comparison of the two-stage EWMA approach with conventional EWMAs for forecasting time series data
- 주제(키워드) Price forecasting , Comparative studies , two-stage exponential weighted moving average model , non-stationary time series , insufficient historical data
- 발행기관 고려대학교 대학원
- 지도교수 한성원
- 발행년도 2018
- 학위수여년월 2018. 8
- 유형 Text
- 학위구분 석사
- 학과 대학원 산업경영공학과
- 세부전공 산업경영공학 전공
- 원문페이지 71 p
- 실제URI http://www.dcollection.net/handler/korea/000000081666
- UCI I804:11009-000000081666
- DOI 10.23186/korea.000000081666.11009.0000818
- 본문언어 영어
- 제출원본 000045953823
초록/요약
This study compared the ability of several Exponential smoothing methods in making forecasts with different kinds of time series data. For this purpose, the 8 scenario time series data are used in the comparison process. Exponential smoothing methods are a basic methodology applied to time series and widely used to forecast the future value of a time series. Nowadays, as the diverse fields of the time series, there are a lot of insufficient historical data when I want to do the forecasts. I find that it is difficult to build a reliable model with the current methods for a satisfactory forecasting accuracy. The recently proposed method, named the two-stage exponential weighted moving average(EWMA) method, compensates this limitation. In our study, I also compared the properties of the two-stage one with other exponential smoothing methods. The forecasting accuracy of each model is measured by examining the prediction error that produced by using the root mean squared error.
more목차
1. Introduction 1
2. Exponential Smoothing Methods 5
2.1. Simple Exponential Smoothing Methods 5
2.2. Double Exponential Smoothing Methods 6
2.3. Triple Exponential Smoothing Methods 6
2.4. Two-stage Exponential Smoothing Methods 8
3. Simulation Study 10
3.1. Simulation Setup 10
3.2. Simulation Results 14
4. Application 40
4.1. Data Preparation 40
4.2. Results 42
5. Conclusion 55
Reference 57
