The purposes of this thesis are to analyze the applicability of diffusion models in Korean situation and to select relatively appropriate sales forecasting model through the comparison of diffusion models with another sales forecasting models.
When applied to Korean electronic products using historical data Bass diffusion model shows good performance in the prediction of peak time and peak sales confirming the results in the past researches.
Dispersion rate model, a modification of Bass diffusion model, shows better explanatory power than Bass diffusion model. Also its predicted peak time and peak sales are closer to actual peak time and peak sales.
To select relatively appropriate sales forecasting model, Bass diffusion model, Box-Jenkins model, Partial Adjustment model, diffusion models with price, diffusion models with income, diffusion models with both price and income are compared on apriori selected econometric criteria. Diffusion model with both price and income is selected as the best sales forecasting model among the one compared.
In conclusion, diffusion models are shown to fit the historical data of Korean electronic products well. Also, they performed better than other compared models. Among diffusion models, the model that considers the price and the income variable simultaneously is shown to be the best diffusion forecasting model.
