3. We have a classification problem with four classes of input vector. Train a perceptron network to solve this problem using the perceptron learning rule. The training set is as follows 222 lo P4 10 Use the following initial weights and biases: b(0) [1