基于Bass与元胞自动机混合模型的快速消费品产品扩散研究

摘要

快速消费品产品扩散主要涉及首次购买和重置购买两个方面.现有的Bass模型对快速消费品的首次购买者有较好的宏观预测能力,而元胞自动机模型可对重置购买者进行较好的预测.借鉴Bass模型和元胞自动机模型各自的优势,提出了一种混合模型,期望用于对快速消费品产品扩散的市场预测.对中国1999年～2006年乳制品的销售量的估计值与实际值的误差为4.11％,拟合度为96.7％,对2007年～2009年乳制品销售量预测值与是价值相差15％,7％和2％,验证了所提出模型的有效性和适用性.%The product diffusion of Fast Moving Consumer Goods ( FMCG) mainly involves two aspects, the initial purchase and the replacement purchase. The current Bass model has the macro forecast ability about initial purchasers, while the cellular automata model can effectively predict the amount of the replacement purchasers. A hybrid model by learning advantages from the Bass and cellular automata models was proposed to forecast the FMCG diffusion. The mean error of estimation values and real values form 1999 to 2006 is 4. 11%, and its fitting degree is 96. 7%; the perdiction values are 15%, 7%, 2% smaller than the real values of Chinese milk sales volumes in 2007,2008,2009 respectively. The experimental results on Chinese milk sales volume verify that the hybrid model is effective and applicable to predict the FMCG product diffusion of Chinese milk sales volumes.