Consider a bipartite network where N consumers choose to buy or not to buy M different products. This paper considers the properties of the logistic regression of the N ×M array of "i-buys-j" purchase decisions, [Yij ] 1≤i≤N,1≤j≤M, onto known functions of consumer and product attributes...