From the fifth-generation mobile communication systems (5G), the architecture of the Internet of Things (IoT) has grown enormously with more promising technologies and connections. However, edge realm components such as sensors and nodes still present a crucial security flaw related to radio frequency attacks. By restricting and stopping information from reaching the proper destination on the IoT's perception layer, jamming - a particular form of a Denial of Service (DoS) attack compromises the availability feature of IoT nodes. Finding and identifying these jammer attacks has been tricky because access to the network from nodes is also broken. To detect and categorize two kinds of jamming in the existing dataset, we propose to utilize three different Shallow Machine Learning (SML) architectures in this paper, including Decision Tree (DT), Support Vector Machine (SVM), and Gradient Boosting Machine (GBM), with the aims of fewer parameter usage, a low power-oriented architecture, and the capacity to handle real-time computer vision tasks. The experimental outcomes depict that SML is a promising approach since it can reach the best detection and classification accuracies (97% and 98.03%, respectively).