Improved Gaussian Grasshopper Optimization Algorithm (IGGOA) and Deep Belief Network (DBN) For Fetal Health Classification

The effective method for continuously monitoring the fetus’s health is called Cardiotocography (CTG). For the past 60 years, CTG has been used in clinical settings, allowing medical professionals to identify early indicators of fetal deterioration. Prior research has solely focused on the classification accuracy of CTG datasets that include the entire feature set, ignoring the classifier computing time. Improved Gaussian Grasshopper Optimization Algorithm (IGGOA) and Deep Belief Network (DBN) are introduced to categorize CTG data. Several factors contribute to the introduction of IGGOA to feature selection: 1) Due to its fragile tail, Gaussian Mutation (GM) is introduced to generate a new offspring closer to the original feature. 2) The development of Levy Flight (LF) has created a more straightforward way to solve the local optimum because grasshoppers have more vital attraction forces;3) OBL has been used to update the present population, and GOA capability is used to determine the FHR classification. In the DBN classifier, CTG samples are categorized as Normal (N), Suspect (S), and Pathological (P). DBN has been analyzed as a Restricted Boltzmann Machine (RBM). Every sub-network hidden layer in the DBN classifier acts as the layer visible to the layer after it. The Softmax function has been used to determine the hidden layer results. The accuracy is about $97.3659 \%$, which is better than other models.