The relationship between users and items, which cannot be recovered by traditional techniques, can be extracted by the recommendation algorithm based on the graph convolution network. The current simple linear combination of these algorithms may not be sufficient to extract the complex structure of user interaction data. This paper presents a new approach to address such issues, utilizing the graph convolution network to extract association relations. The proposed approach mainly includes three modules: Embedding layer, forward propagation layer, and score prediction layer. The embedding layer models users and items according to their interaction information and generates initial feature vectors as input for the forward propagation layer. The forward propagation layer designs two parallel graph convolution networks with self-connections, which extract higher-order association relevance from users and items separately by multi-layer graph convolution. Furthermore, the forward propagation layer integrates the attention factor to assign different weights among the hop neighbors of the graph convolution network fusion, capturing more comprehensive association relevance between users and items as input for the score prediction layer. The score prediction layer introduces MLP (multi-layer perceptron) to conduct non-linear feature interaction between users and items, respectively. Finally, the prediction score of users to items is obtained. The recall rate and normalized discounted cumulative gain were used as evaluation indexes. The proposed approach effectively integrates higher-order information in user entries, and experimental analysis demonstrates its superiority over the existing algorithms.

As the popularity and dependence on the Internet increase, DDoS (distributed denial of service) attacks seriously threaten network security. By accurately distinguishing between different types of DDoS attacks, targeted defense strategies can be formulated, significantly improving network protection efficiency. DDoS attacks usually manifest as an abnormal increase in network traffic, and their diverse types of attacks, along with a severe data imbalance, make it difficult for traditional classification methods to effectively identify a small number of attack types. To solve this problem, this paper proposes a DDoS recognition method CVWGG (Conditional Variational Autoencoder-Wasserstein Generative Adversarial Network-gradient penalty-Gated Recurrent Unit) for unbalanced data, which generates less noisy data and high data quality compared with existing methods. CVWGG mainly includes unbalanced data processing for CVWG, feature extraction, and classification. CVWGG uses the CVAE (Conditional Variational Autoencoder) to improve the WGAN (Wasserstein Generative Adversarial Network) and introduces a GP (gradient penalty) term to design the loss function to generate balanced data, which enhances the learning ability and stability of the data. Subsequently, the GRU (Gated Recurrent Units) are used to capture the temporal features and patterns of the data. Finally, the logsoftmax function is used to differentiate DDoS attack categories. Using PyCharm and Python 3.10 for programming and evaluating performance with metrics such as accuracy and precision, the results show that the method achieved accuracy rates of 96.0% and 97.3% on two datasets, respectively. Additionally, comparison and ablation experiment results demonstrate that CVWGG effectively mitigates the imbalance between DDoS attack categories, significantly improves the classification accuracy of different types of attacks and provides a valuable reference for network security defense.