Research on Identification of Multi-Source Heterogeneous Information Risks: An Integrated Deep Random Forest and Markov Games Method

This research proposes an integrated deep random forest (RF) and Markov games method for identifying multi-source heterogeneous information risks to ensure system security against multi-source heterogeneous information and provide an effective auxiliary tool for information risk defense. It adopts packet capture to obtain multi-source heterogeneous information as the object for risk identification, analyzes its information vulnerability, and employs the deep RF algorithm and Markov games to detect the threat risk and attack risk of multi-source heterogeneous information. The entire risk identification workflow involves the derivation of information risk values and the determination of risk types using the integrated deep RF and Markov games method. With both static and dynamic experimental environments considered, the optimized method reduces the information risk identification error rate by approximately 13 % and the risk value identification deviation by approximately 0.05 compared with two traditional risk identification methods, demonstrating superior performance.