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