Performance of the perceptron algorithm for the classification of computer users

The perception algorithm classify computer users. Test data was used to was collected interactively-from 5 users over a 5 week period. The times between keystrokes entered in a password formed the measurement vector. Decision functions were derived using part of the data (training data) to compute the weight vectors. The decision functions were applied to the remaining data (testing data) to classify the users. Four users were classified correctly with no error, and one user was misclassified with a 10IZO error resulting in an overall misclassification error of 270.