Classification by cascaded threshold elements

摘要

A method is presented to synthesize threshold element networks by iterative computation of the parameters. The modification of the parameters, which is under the control of a gain function, is carried out sequentially. The training procedure of the whole cascade is subdivided into separate training periods during which only the parameters of one threshold element are modified. The algorithm starts with one threshold element. If a given pattern set is not linearly separable the structure of the classifier is extended automatically to a general cascade. Computer simulations indicate good results.