A STATISTICAL ANALYSIS OF THE ENGINEERING APPROACH TO NAVY SHIPBUILDING COST ESTIMATION
GEORGE WASHINGTON UNIV WASHINGTON DC PROGRAM IN LOGISTICS
Pagination or Media Count:
The paper discusses the feasibility of developing regression models to predict the total cost of a Navy ship using the physical weights of the ship components as independent variables was investigated. The various forms of regression analyses fall under the following three categories Linear multiple regression analysis Non-linear multiple regression analysis and Adding-up process, which is an aggregation of two-variable regression analyses. It was found that the linear model is preferable over the non-linear model and the adding-up process. If the samples are properly selected, linear models which are statistically significant can be derived. Given its superiority over the other two models, the degree of accuracy of the linear model is still not high enough to produce a dependable point estimation for the total cost of the ship.
- Statistics and Probability
- Marine Engineering