VPC

Voted perceptron classifier

    W = VPC(A)
    W = VPC(A, N)

Input
A	Dataset
N	Number of sweeps

Output
W	Voted perceptron classifier

Description

The classifier trains an ensemble of perceptrons on dataset A. The  training procedure performs N full sweeps through the training data. If N is not specified, 10 sweeps are performed. The total number of  perceptrons in the ensemble is equal to N x GETSIZE(A, 1).

The classification of new objects is performed by allowing the ensemble  of perceptrons to vote on the label of each test point.

Reference(s)

Y. Freund and R.E. Schapire. Large Margin Classification Using the Perceptron Algorithm. Machine Learning 37(3):277-296, 1999.

See also

(prtools, guide), datasets, mappings, perlc, adaboostc,