Hyperparameter Optimization in CNN: A Review
Anu Sharma, Dharmender Kumar · 2023
Hyperparameter optimization is an important issue in convolutional neural networks (CNNs), which is an appropriate deep learning network for image classification. Several classical and metaheuristic algorithms are often employed to optimize hyperparameters. In the present paper, different algorithms are compared by using various evaluation measures. The literature reveals that genetic algorithms (GA) and PSO or particle swarm optimization are the most effective algorithms for hyperparameter optimization used in a variety of domains including disease diagnosis; face detection, and handwritten character recognition. Further, it is concluded that optimized CNNs, are essential to the early detection of a variety of diseases, which can assist physicians and clinicians in saving lives by providing accurate and timely predictions.