Multiple Data Stream Analysis and Short Term Forecasting
Abstract
The research is concerned with the development of techniques for the analysis and interpretation of event data for forecasting, using quantitative historical analogies. The work builds on prior research on the coding of event data and the development of quantitative indicators for defense analysis. The analysis is based on event data coded from the daily 'New York Times' and 'The Times' of London from mid-1971 to mid-1974. The analysis illustrates the similarities and differences in the international interests, attention and viewpoints of the US and UK as reflected in these public media sources. A short term forecasting technique is developed based on historical patterns derived from univariate time series analysis. The use of these patterns and their subsequent outcomes as analogues for forecasting the outcomes of current situations is argued and illustrated. In addition, a research design for multivariate analysis is described.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1975
- Accession Number
- ADA008172
Entities
People
- Theodore J. Rubin