本研究將應用模糊逐段羅吉斯成長模型、Gompertz模型、傳統羅吉斯成長模型及Norton與Bass模型預測動態隨機存取記憶體(Dynamic Random Access Memory,DRAM)產業八個世代的市場成長趨勢,結果顯示只有模糊逐段羅吉斯成長模型及Norton與Bass模型能夠有效的預測DRAM產業在市場上複雜世代的成長趨勢。此外,模糊逐段羅吉斯成長模型偵測的改變點也可以合理的解釋DRAM產業在當時發生的現象與世代交替情形並指出影響市場情形的概況,因此模糊逐段羅吉斯成長模型明顯是一個較健全與具有更好預測能力的方法。
The fuzzy piecewise logistic growth model, Gompertz model, traditional logistic growth model, and Norton and Bass model are applied to predict the growth trends of eight generations of the DRAM industry market. The results show that only the fuzzy piecewise logistic growth model and the Norton and Bass model can effectively forecast the growth trends of complex generations in the DRAM industry. In addition, the change points detected by the fuzzy piecewise logistic growth model can also be used to reasonably explain the phenomenon and the status of generation replacement in the DRAM industry at particular times, and can indicate the overview of an influenced market situation. Therefore, the fuzzy piecewise logistic growth model is more robust and has a greater prediction capability.