Optimization of Machine Learning Algorithms Through Hyper-parameter Tuning Applied for the classification of Arabic news

Abstract In the field of machine learning, hyperparameter tuning is a technique used to determine which learning algorithms have the most effective parameters. Using the Grid Search strategy as part of the machine learning algorithms, we suggest a few different ways to improve the accuracy of text categorization. In this research, we combined the TF-IDF feature selection approach with three different machine learning algorithms. These algorithms are Multinomial Logistic Regression MLR, Support Vector Machine SVM, and Artificial Neural Network ANN. According to the results of our tests, hyperparameter adjustment can significantly improve the classifier's performance.