Abstract
In this study, a change in the number of multi-functional forestry machinery in Japan was predicted using growth-curve models. Logistic-curve and Gompertz-curve were fit into the number of machines during fiscal year 1988-1997, and the AIC of both models were compared for validity evaluation. The model parameters were estimated by nonlinear regression analysis (maximum likelihood method). As a result of estimation, the predicted level of machine saturation was 2,000 by Logistic-curve model and 2,900 by Gompertz-curve model. Gompertz-curve model was chosen as a proper model because the AIC was fewer. The following attributes were expected to increase; "harvester" in machine type, "Kanto and Chubu" and "Kinki, Chugoku and Shikoku" in region, "other cooperatives (support centers for the security of forest labor etc.)" for machine utilization in ownership. All in all, the rate of machine increase was predicted to level off, so we should make more efforts to utilize machines effectively than to increase in number.
