COMMUNITY DETECTION ON SOCIAL NETWORKING META MODEL BASED DATA ( CASE STUDY : E-COMMERCE (AMAZON) AND ONLINE REFERENCE MEDIA (DBLP) )

• Main Author: Yulian Pratama - NIM. 23216019 , Nano
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/29543
• Institution: Institut Teknologi Bandung

<p align="justify">The enormous availability of data due to the impact of social media presence, ecommerce, and online reference media is inevitable. Facebook, Amazon, and dblp are the real examples of the three domains. The data generated from the three domains are often being used by researchers and interested parties to be able to mine valuable information through analysis on related data that represented in a graph. The process of analyzing related entities through graph theory or network is known as Social Network Analysis (SNA). Before conducting an SNA, generally require a model that can be used as a guide to perform data collection and analysis to be performed. <br /> <br /> Having a data model that matches directly with the domain model to be analyzed will help in better understanding the data, communicate effectively, and avoid unnecessary work. Until now, in the case of dblp and Amazon there has been no research focusing on graph-based data modeling that can be used to support the needs of SNA. <br /> <br /> Social networking meta-model is a model which is designed based on the abstraction of Facebook to be used as a guidance in helping to build graph-based data model on other social networking such as LinkedIn, Youtube, Instagram, and Twitter. The resulting data model can be used to support the related analysis needs in each of the social networking. Therefore, this research tries to answer whether a meta-model of social networking can be implemented to the others domains than social media especially in Amazon and dblp. <br /> <br /> Based on the analysis and design conducted, the social networking meta-model can be used to generate graph data model from Amazon and dblp. Several test are also conducted on both graph data models that have been designed. Community detection analysis is used as a case study on submodels of graphs that have been produced on Amazon and dblp to demonstrate that the designed graph model can be used to support the SNA process. At the end of this research, it is concluded that the meta-model of social networking can be implemented to Amazon and dblp. The resulting model can be implemented to support community detection analysis.<p align="justify"> <br />