Anomaly detection is critical across domains, from cybersecurity to fraud prevention. Graphs, adept at modeling intricate relationships, offer a flexible framework for capturing complex data structures. This paper proposes a novel anomaly detection approach, combining Graph Convolutional Networks (GCNs) and Density-Based Spatial Clustering of Applications with Noise (DBSCAN). GCNs, a specialized deep learning model for graph data, extracts meaningful node and edge representations by incorporating graph topology and attribute information. This facilitates learning expressive node embeddings capturing local and global structural patterns. For anomaly detection, DBSCAN, a density-based clustering algorithm effective in identifying clusters of varying densities amidst noise, is employed. By defining a minimum distance threshold and a minimum number of points within that distance, DBSCAN proficiently distinguishes normal graph elements from anomalies. Our approach involves training a GCN model on a labeled graph dataset, generating appropriately labeled node embeddings. These embeddings serve as input to DBSCAN, identifying clusters and isolating anomalies as noise points. The evaluation on benchmark datasets highlights the superior performance of our approach in anomaly detection compared to traditional methods. The fusion of GCNs and DBSCAN demonstrates a significant potential for accurate and efficient anomaly detection in graphs. This research contributes to advancing graph-based anomaly detection, with promising applications in domains where safeguarding data integrity and security is paramount.

As various types of data are generated on the social Internet of things (SIoT), which combine the Internet of things (IoT) and social networks, the relations of IoT devices should be established for necessary data exchange. In this paper, we propose a user recommendation scheme that facilitates data sharing through an analysis of an interaction between an IoT device and a user in the SIoT. An interrelation between a user and an IoT device as well as an interrelation between users exist simultaneously in the SIoT. Hence, the interaction between users must be analyzed to identify the interest keywords, and the interaction between IoT devices and users to determine the user’s preference of IoT device. Moreover, the proposed scheme calculates the similarity between users based on the IoT device preference based on IoT device usage frequency and interest keywords, which are identified through an analysis between the user and IoT device and that between users. Subsequently, it recommends top-N users who have a high similarity as the users for data sharing. Furthermore, the performance of the proposed scheme is verified through performance evaluation based on the precision, recall, and F-measure.

As graph stream data are continuously generated in Internet of Things (IoT) environments, many studies on the detection and analysis of changes in graphs have been conducted. In this paper, we propose a method that incrementally detects frequent subgraph patterns by using frequent subgraph pattern information generated in previous sliding window. To reduce the computation cost for subgraph patterns that occur consecutively in a graph stream, the proposed method determines whether subgraph patterns occur within a sliding window. In addition, subgraph patterns that are more meaningful can be detected by recognizing only the patterns that are connected to each other via edges as one pattern. In order to prove the superiority of the proposed method, various performance evaluations were conducted.

As a large amount of stream data are generated through sensors over the Internet of Things environment, studies on complex event processing have been conducted to detect information required by users or specific applications in real time. A complex event is made by combining primitive events through a number of operators. However, the existing complex event-processing methods take a long time because they do not consider similarity and redundancy of operators. In this paper, we propose a new complex event-processing method considering similar and redundant operations for stream data from sensors in real time. In the proposed method, a similar operation in common events is converted into a virtual operator, and redundant operations on the same events are converted into a single operator. The event query tree for complex event detection is reconstructed using the converted operators. Through this method, the cost of comparison and inspection of similar and redundant operations is reduced, thereby decreasing the overall processing cost. To prove the superior performance of the proposed method, its performance is evaluated in comparison with existing methods.