| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 384420 | Expert Systems with Applications | 2012 | 8 Pages |
In this study, we propose a product network analysis, a network-based analysis to analyze a network-leveled relation among all products. Compared to market basket analysis, which focuses on the transaction-leveled relation between products, the suggested product network analysis focuses on extended network-leveled point of view of the relation between all products. For such a purpose, we suggest two kinds of product networks, market basket networks and co-purchased product networks. Two networks are comparatively evaluated to analyze the topological characteristics and the structure of those networks. The extended use of market basket analysis, network-leveled analysis are expected to be more effectively and efficiently used in personalized services, such as cross selling, up selling, and personalized product display utilizing the deep relation between products.
► Co-purchased product networks and market basket networks are suggested to expand the concept of market basket analysis. ► Two networks are evaluated to analyze the topological characteristics of networks. ► We found that bestsellers do not necessarily occupy centrality in network. ► Products purchased with other ones all the time are not necessarily bestsellers. ► And products with higher centrality are found to be effective in sales promotion and increase sale volumes.
