In part A of this assignment, we are given the Cardiotocography dataset and tasked to design a multilayer perceptron network to predict one of the three fetal states (N: Normal, S: Suspect, and P: Pathologic) given a record of data. Each data record has 23 fields. While there are multiple labels available in our dataset, the main interest of this assignment was placed on building a predictive model for the NSP label. Thus, we have 21 features/attributes and 1 label to work with in the designing and evaluation of this model. The results of this experiments shown how existing neural networks’ performances may be improved by: • Tuning the hyperparameters of the model • Modifying the model architecture.