AHP-Weighted Conversion of Anomaly Detection Datasets for Network Security Evaluation

The challenge of network security situation prediction lies in the lack of a unified time-series dataset and the influence of multiple security factors, making it difficult to reach consensus at the foundational stage. To address this issue, we designed a process to transform anomaly detection datasets into network security sequence datasets and proposed a method for setting security factor weights based on the Analytic Hierarchy Process (AHP). This method combines subjective and objective approaches to effectively determine the weights of various security factors, thereby obtaining a comprehensive security situation value through the weighted sum of network security elements and their weights. We conducted experiments on the constructed datasets using the commonly employed LSTM neural network architecture, validating the effectiveness of the proposed AHP-based security factor weighting method outlined in this paper.