Enhancing Data Quality by Identifying Influential Nodes: Integrating Complementary Features with CoCoFISo

With the growth in the usage of social networks, data quality improvement has become important to optimize and speed up various methodologies. One approach to achieve this aim is through the removal of unnecessary data. This paper discusses this issue by ranking nodes based on their potential for dissemination. Earlier approaches to detecting influential nodes have been either manual combinations of structural graph properties or MCDM (Multiple Criteria Decision Making) methods. Yet, there is still a need for a method with improved accuracy and resolution. This paper emphasizes the necessity of choosing complementary features and the optimal ways of combining them. To attain this objective, we employ five various graph structural properties, including K-Shell (KS) and PageRank (PR), that are complementary. This study is the first to employ CoCoFISo (Combined Compromise for Ideal Solution) in the Multi-Criteria Decision Making (MCDM) method to obtain the key nodes. The assessments are made regarding accuracy, correlation, computational complexity, and resolution on five real-world networks. The experimental results show that the proposed method outperforms 12 state-of-the-art methods, including HCURank and HGSM.