Selection and Classification Using a Forecast Applicant Pool.
Abstract
A time series analysis model was developed to forecast the quality (i.e., means and standard deviations of aptitude scores) and quantity (i.e., total number of applicants) of the Air Force's future applicant pool. By forecasting future talent of applicants and their number, an approximate optimal assignment solution could be obtained even though the applicants have to be assigned one at a time. The model was developed on 258,588 subjects who had taken the Airman Qualifying Examination during 1971 to 1974. The model included trend and seasonal components and was evaluated by applying it to forecast monthly means, standard deviations, and total number of applicants. Results in terms of mean absolute deviation and squared absolute deviation scores indicated that the model could forecast within one point of the actual observed score values across two years for means and standard deviations, but not as well for total number of applicants.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1976
- Accession Number
- ADA025327
Entities
People
- William H. Hendrix
Organizations
- Brooks Air Force Base