基于SOM的市场细分研究

本文应用数据挖掘技术对市场细分的研究包括:如何利用SOM聚类技术解决市场细分问题;如何将SOM聚类的市场细分结构结果可视地呈现给市场决策人员的问题。这两部分研究整合为企业市场战略集成一个研究途径,包括:偏好数据的收集与前处理步骤、偏好数据聚类步骤、偏好数据的可视化步骤。实验结果表明,本文提出的研究途径成功发现了人工数据集中预设的聚类模式,与通常研究途径相比具有明显的优点。在实际市场数据分析中,获得了与事实相符的结论并提供了有价值的决策支持信息。

SOM-Based Market Segmentation Research

The research in market segmentation includes two main parts.We focus firstly on discussing the market seg- mentation problem by applying SOM clustering technique in data mining discipline.The second part is focus on displa- ying market segmentation structure.We apply visualization technique to represent the market structure clearly in a two dimensional plane so that the marketers can make their market strategies easier.The two main parts are organized as an integrated approach.Such an approach includes three core steps:preference data collecting step,preference data clus- tering step by SOM neural networks and visualization step by ideal point model.The experiments show that the ap- proach yields meaningful results and is comparable and complementary to the most general ones.