Performance evaluation and design of multiple time series based forecasting systems Public Deposited

http://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/zk51vm31t

Descriptions

Attribute NameValues
Creator
Abstract or Summary
  • The study evaluates performances of three multiple time series (MTS) forecasting methods; investigates possible improvements in MTS forecasting operations, and proposes a multiple time series based forecasting system. Time series considered are finite, linear, covariance stationary, and discrete. Specific objectives for the thesis were: (a) conducting comparative studies of MTS estimation methods; (b) investigating the effectiveness of composite forecasts in MTS analysis; (c) examining performance of process control procedures in MTS forecasting operations; and (d) proposing an overall approach to MTS estimation, model building and forecasting procedures. Multivariate exponential smoothing, multivariate Yule-Walker equations, and Chitturi's discounted least squares parameter estimation methods were used in the study. Time series schemes which are most suitable for using these estimation methods are multiple autoregressive processes, MAR(s,p); and mixed autoregressivemoving average processes, MARMA(s, 1, 1). Comparative studies were conducted on selected real economic time series. Outcomes of forecasting were measured in terms of weighted mean square errors. Computational efforts were recorded and compared. Proper data treatments were taken to ensure desired characteristics such as stationarity, and nonseasonality. The techniques of composite forecasting were studied and applied to MTS analysis. Performance of all four alternatives for composite forecasts were reviewed and favorable results were reported. In an effort to reduce computation expenses in MTS analysis, process control procedures were implemented. A trivariate MTS model was then used to demonstrate the application of process control procedures in MTS forecasting operations. An overall approach to MTS estimation, model building and forecasting procedure has been developed. It is a two-phase procedure. The estimation phase is a multi-variable analogue of the Box-Jenkins iterative process for model building in single time series analysis. Process control procedures and composite forecasting techniques are incorporated in the second phase. The complete process of the proposed multiple time series based forecasting systems is illustrated through a trivariate MTS model.
Resource Type
Date Available
Date Copyright
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Committee Member
Academic Affiliation
Non-Academic Affiliation
Subject
Rights Statement
Peer Reviewed
Language
Digitization Specifications
  • File scanned at 300 ppi (Monochrome) using ScandAll PRO 1.8.1 on a Fi-6770A in PDF format. CVista PdfCompressor 5.0 was used for pdf compression and textual OCR.
Replaces
Additional Information
  • description.provenance : Submitted by Georgeann Booth (gbscannerosu@gmail.com) on 2013-10-30T23:14:48Z No. of bitstreams: 1 ChenKuei1977.pdf: 834501 bytes, checksum: da9c7f0493cb5a351e49bde4b587fcb8 (MD5)
  • description.provenance : Approved for entry into archive by Deborah Campbell(deborah.campbell@oregonstate.edu) on 2013-11-04T22:51:13Z (GMT) No. of bitstreams: 1 ChenKuei1977.pdf: 834501 bytes, checksum: da9c7f0493cb5a351e49bde4b587fcb8 (MD5)
  • description.provenance : Approved for entry into archive by Patricia Black(patricia.black@oregonstate.edu) on 2013-11-01T16:36:56Z (GMT) No. of bitstreams: 1 ChenKuei1977.pdf: 834501 bytes, checksum: da9c7f0493cb5a351e49bde4b587fcb8 (MD5)
  • description.provenance : Made available in DSpace on 2013-11-04T22:51:13Z (GMT). No. of bitstreams: 1 ChenKuei1977.pdf: 834501 bytes, checksum: da9c7f0493cb5a351e49bde4b587fcb8 (MD5) Previous issue date: 1976-12-03

Relationships

Parents:

This work has no parents.

Last modified

Downloadable Content

Download PDF

Items