System and method of forecasting print job related demand

A system of forecasting print-related demand in a document production environment, comprising:

• (A) a processor;
• and
• (B) a processor readable storage medium in communication with the processor, the processor readable storage medium containing one or more programming instructions for:
• (1) providing a hidden layer including a plurality of neurons and causing each one of the plurality of neurons to be weighted with a first set of print-related demand data,
• (2) providing an input layer including a plurality of inputs, the plurality of inputs communicating with the plurality of neurons of the hidden layer,
• (3) communicating a second set of print-related demand data to the plurality of inputs of the input layer, the second set of print-related demand data corresponding with a demand series obtained from the document production environment, and
• (4) responsive to communicating the second set of print-related demand data to the plurality of inputs, a print-related demand forecast output for the document production environment is generated with the plurality of weighted neurons of the hidden layer.

A method for optimizing a print-related demand forecasting system, comprising:

• (A) providing a neural network, the neural network including a hidden layer with a plurality of neurons;
• (B) setting the plurality of neurons to a first number;
• (C) using a set of print-related demand data to obtain a print-related demand forecast with the neural network when the plurality of neurons is set to the first number;
• (D) measuring a first level of accuracy associated with the print-related demand forecast obtained with (C);
• (E) setting the plurality of neurons to a second number;
• (F) using the set of print-related demand data to obtain a print-related demand forecast with the neural network when the plurality of neurons is set to the second number;
• (G) measuring a second level of accuracy associated with the print-related demand forecast obtained with (F);
• and
• (H) when a selected condition is met, optimizing the hidden layer with one of the first number of neurons and the second number of neurons on the basis of said measuring of (D) and (G).