COMMUNITY AND IMPORTANT ACTORANALYSISWITH DIFFERENT KEYWORDS IN SOCIAL NETWORK (CASE STUDY: CYBERTERRORISM IN TWITTER)

• Main Author: CAHYANA (NIM : 23215366), NANANG
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/23428
• Institution: Institut Teknologi Bandung

Analysis of groups and actors on social networks usually uses community detection and centrality methods. Community detection method aims to detect groups that exist on social networks, and centrality aims to find the most important actors in social networking groups. Data input for community detection and centrality is a large and structured input data. In the case study of Twitter, the input data for community detection and centrality is derived from the crawling data, which also requires a keyword input. Most research on crawling data Twitter uses only one keyword so that the resulti community detection and centrality results have not depicted the strong groups and actors in the graphs. <br /> <br /> <br /> For that, a new method is needed to produce strong groups and actors. One of the new methods, namely the use of different keywords but has a one topic discussion. To know the value of relationships of different keywords then used the fuzzy relation method so that different keywords are said to have relationships between keywords. Different and related keywords are the input to the community detection and centrality methods resulting in various groups and actors. To find out influence groups and actors generated between the use of one keyword with different keywords and berelasi then merged group grafts. Merging group graphs from the results of each different keyword based on the similarity of actors in each of the groups generated by different keywords. In the end, we get the value of group graphs and actors based on the use of different and related keywords <br /> <br /> <br /> The results of this study obtained that the use of different keywords and berelasi can produce the group network and the most important actor stronger and changes in the value of group graphs and actors. Changes in the value of group graphs in the form of vertex, edge, average geodesic distance, and modularity values that have increased but decreased in density. While the change in the value of actors in the form of degree, betweenness, and eigenvector centrality values that experienced an increase but decreased value in closeness centrality. <br /> <br /> <br /> The implication of this research is the existence of a simple method that produces the most important groups and actors that are getting stronger based on different keywords and relate on Twitter's social network.