artificial intelligence

A Problem 2 Device under test DUT V An electrical device is experiencing some failures. It has been noticed that we may be able to detect the failure by monitoring two features: the current (1) and the voltage (V) across the device. We would like to develop a perceptron (linear classifier) to detect the fault from these two features. (To unify the answers, try to use I and V as subscripts for features, weights, etc.) We have performed the following set of experiments and labeled the results, where (+) means that a failure 5 was observed: 4 Exp # V 1 Result Failure? 3 1 2 2 2 3 + 2 3 3 3 4 1 1 5 4 1 11 N AwNN + - V + 0 1 2 3 4 5 d) We would like to train the perceptron using the training examples. Fill out the following tables with your steps starting from the initial weight shown. The weight elements are ordered such that they correspond to: [Wies, wv,w). The steps are also ordered starting from data point #1 till data point #5. step Weight vector Score Decision Correct? Training action [1,1,1) w=w + (-1)"[1,2,2) [0,-1,-1] 5 +1 no 1 2 4 un w=1 Final Weights e) Did the final weight vector (after training) reach to a satisfactory value? In other words, is it now able to classify all training data points correctly? If not, what can we do to get a better classifier? f) Sketch the classifier decision line before and after training and indicate the decision regions. Add the training data points to your sketch. 5- 5 5 5 4 4 3 3- N N 1 1 - V 0+ 0 1 2 3 4 5 Initial classifier (using w=(2,0,01) 0 + 0 1 2 3 4 5 Final classifier (using final w after training)