The brain is the organ that controls the activities of all parts of the body. The tumor is familiar as an irregular outgrowth of tissue. Brain tumors are an abnormal lump of tissue in which cells grow up and redouble uncontrollably. It is categorized into different types based on their nature, origin, growth rate, and stage of progress. Detection of the tumor by traditional methods is time-consuming and does not widen to diagnose a large amount of data and is less accurate. So, the automatic diagnosis of the tumors in the brain by magnetic resonance imaging (MRI) plays a very important role in computer-aided diagnosis. This paper concentrates on the diagnosis of three kinds of brain tumors (a meningioma, a glioma, and a pituitary tumor). Machine learning algorithms: KNN, SVM, and GRNN are suggested to increase accuracy and reduce diagnostic time by using a publicly available dataset, features that are extracted of images, data pre-processing methods, and the principal component analysis (PCA). This paper aims to minimize the training time of the suggested algorithms. The dimensionality reducing technique is applied to the dataset and diagnosis using machine learning algorithms, such as Support Vector Machines (SVM), K-Nearest Neighbor (KNN), and Generalized Regression Neural Networks (GRNN). The accuracies of the algorithms used in diagnosing tumors are 97%, 96.24%, and 94.7% for KNN, SVM, and GRNN, respectively. The KNN is therefore regarded as the algorithm of choice.
Proton Exchange Membrane Fuel Cells (PEMFC) have several merits over the other fuel cells like low operating temperature, high power density, light weight, and robustness, fitting them in a wide variety of green energy applications. The most important application area of the PEMFC is hydrogen cars (Fuel Cell Electric Vehicles – FCEV), stationary and portable hydrogen power systems. Throughout the presented research work the complete three cycles of the fuel cell control are presented which are the control system, the FCEV control and the AI based metaheuristic optimization. The presented algorithm includes the PI controller – PI-based HHO-PI-based AEO-PI-based GWO-PI-based PSO. An objective statistical comparison among the presented algorithms to select the best one of them according to perfect criteria.