QUESTION 4 a) Identify two (2) advantages of neural networks and its applications that make it more preferable in real world. (4 marks) b) State one (1) characteristic of the data needed in the standardization process for neural network. (2 marks) c) In a neural network, knowing the weight and bias of each neuron is one of the most important steps. What is the role of weights and bias? (2 marks) d) In training a neural network, Diagram 1 indicates that the error rate does not decrease during the starting of the first few of the iterations. Identify one (1) potential problem for this situation. Error rate Iteration Diagram 1 (1 mark)