The Edge computing streamlines data processing by bringing it to the edge device, reducing response time, saving bandwidth, and enhancing application efficiency. As edge computing gains prominence, the challenges related to data outsourcing, security, and privacy become increasingly significant. Our goal is to enhance current methodologies by addressing privacy concerns and proposing performance-enhancing strategies for data outsourcing. The proliferation of edge devices in network environments poses a challenge for conventional intrusion detection systems (IDSs) due to extensive data processing. The evolving architecture of edge computing, with its layered structure, complicates privacy and security concerns, making intrusion detection in decentralized systems challenging. This study addresses this by proposing an enhanced EC-IDS, employing multiple machine learning models and optimized feature selection methods with filter-based SMOTE technique. The completed work on the UNSW-NB15 dataset demonstrates the proposed ECIDS-DT superior performance with respect to previous state of art.