A Combined Regression and Box-Jenkins Prediction Model for Reenlistment in Selected Navy Ratings.
Abstract
This thesis examines five Navy ratings using two distinct modeling techniques in an effort to predict first term reenlistments. The techniques utilized are Box-Jenkins time series analysis and linear regression. A combined model utilizing both techniques is also developed. The ability of the models ot predict is considered adequate for three of the five ratings and not adequate for two of the ratings. The regression models utilizing 20-24 year old unemployment as the only independent variable yielded surprisingly good predictions, once the time series patterns in the data were modeled. Yeoman, Storekeepers, Operations specialists, electronics technicians and boiler technicians were used in models. Keywords include: Time series; Box-Jenkins methodology; Univariate analysis.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1984
- Accession Number
- ADA151106
Entities
People
- K. J. Sherry
Organizations
- Naval Postgraduate School