Competition State Visualization for Sales Record Mining

摘要

To determine the competition state among products, this paper is used to present a visualization technique for mining interesting knowledge. First, the technique determines the relations among products from volumes of sales records. Then, it identifies the preference transition as a direction for the identified relations. According to the type of direction and the number of connected links, the technique assigns the size and color of a product a node and generates a graph. Finally, this paper describes the semantics of the competition state among the products in a graph. A numerical example of the sales records of 172 kinds of products and their 90,000 sales records is also illustrated.