A Data-Driven Method for DDoS Attack Detection and Effect Determination

With the in-depth penetration of the Industrial Internet into key fields such as energy and transportation, DDoS attacks have become a major security risk for the Industrial Internet due to their high frequency and wide range of damage. Currently, there are problems in the detection and effect evaluation of DDoS attacks, including feature redundancy, subjective weight assignment, and inconsistent evaluation standards. This paper proposes a data-driven lightweight method: random forest is used to screen a subset of core features with high discriminative power, the entropy weight method is adopted for objective weight assignment to construct a weighted matrix, and the TOPSIS algorithm is applied to calculate the attack index to realize attack detection and effect evaluation. Experiments show that this method achieves an accuracy rate of over 95% in detecting different types of DDoS attacks, the attack index can accurately reflect the attack effect, and it reduces computational overhead, making it suitable for industrial scenarios.