Research on Identification Methods for False or Unrelated Information in Network Resource Content

There are problems in the recognition of false or irrelevant information in network resource content, such as low recognition rate and high time cost. Therefore, this article proposes a new method for identifying false or irrelevant information in network resource content. By analyzing the network environment in which network resource content is located, starting from the intelligent agent node network, the local topology structure of the network environment is determined. Gaussian distribution and covariance matrix are used to detect whether false or irrelevant information in network resource content is injected into the network, and the maximum standard residual method is introduced to reduce the identification error of false or irrelevant information. Introducing the PCFG analysis tree to analyze the deep syntactic features of false or irrelevant information in network resource content, linking the probability of nodes having false or irrelevant information in the network with the number of nodes in the same cluster among real information nodes, and using the evolutionary equilibrium function to iteratively and recursively extract feature values to achieve feature extraction of false or irrelevant information in the network; By using network node state estimation values and average filtering approximation methods, the trust degree of network nodes in network resource content is calculated to confirm that false and irrelevant information falls into the confidence region of network nodes. Based on this, the density peak of shared neighbors is used to cluster false or irrelevant information of network resources, and the predicted value of state estimation is introduced as the recognition basis. A fixed recognition threshold and confidence function are set to achieve recognition research. The experimental results show that the proposed recognition algorithm performs well in recognition rate, with an average recognition rate of 93.5%, an average time cost of 122 s, and a network stability coefficient of 0.94 or above, reaching a maximum of 0.96, fully verifying the effectiveness and reliability of the method.