Observation of network structure in Amazon.com

Abstract

Amazon.com is among the largest bookstores on the internet. It provides the sales rank of each book. Our hypothesis is as follows: if a book has low sales rank (i.e., it is well-sold), the related book also has a low sales rank. In the small world principle, if a network is connected, any two nodes are connected with relatively small number of links. If this hypothesis is true, any book can link to a bestseller book with small number of distance by linking some related books. In this paper, we design an algorithm to ascertain our hypothesis, and analyze the network structure of Amazon.com.