Today, securing devices connected to the internet is challenging as security threats are generated through various sources. The protection of cyber-physical systems from external attacks is a primary task. The presented method is planned on the prime motive of detecting cybersecurity attacks and their impacted parameters. The proposed architecture employs the LYSIS dataset and formulates Multi Variant Exploratory Data Analysis (MEDA) through Principle Component Analysis (PCA) and Singular Value Decomposition (SVD) for the extraction of unique parameters. The feature mappings are analyzed with Recurrent 2 Convolutional Neural Network (R2CNN) and Gradient Boost Regression (GBR) to identify the maximum correlation. Novel Late Fusion Aggregation enabled with Cyber-Net (LFAEC) is the robust derived algorithm. The quantitative analysis uses predicted threat points with actual threat variables from which mean and difference vectors are evaluated. The performance of the presented system is assessed against the parameters such as Accuracy, Precision, Recall, and F1 Score. The proposed method outperformed by 98% to 100% in all quality measures compared to existing methods.