International Journal of Technology And Business
Vol 2 No 2 (2018): IJTB | INTER1NATIONAL JOURNAL OF TECHNOLOGY AND BUSINESS

SOCIAL NETWORK ANALYSIS TOWARD TWITTERS USERS AGAINST HOAX IN INDONESIA WITH SINGLE CLUSTER MULTI-NODE METHOD USING APACHE HADOOP HORTONWORKSTM DISTRIBUTION WITH INTACT-GROUP COMPARISON EXPERIMENT

Arini, Arini (Unknown)



Article Info

Publish Date
10 Nov 2018

Abstract

Social network analysis (SNA) examines the relationship of nodes in the graph. This study uses Hadoop technology with the use of clustering method (High-Performance Cluster) and 5 SNA parameters on hoax topic, by manipulating the iteration and assign the value of t-max on Gephi tools. Based on 1 month of observations, the authors found 18 days where keyword hoax became the trending topic on Twitter. This means 58% of Twitters users are more often talk about hoaxes. Intact-group comparison experiments are applied to Twitter data that will be grouped into 2 groups (controls and experiments). The authors were able to get data as much as 16,400 data during data gathering phase. Variable changes were performed in the experiment group. The addition of iteration from 100 to 200, which increase the value of degree centrality from 3601 to 5070 or about 40.79%. Assignment of t-max from none to 60s increases 33.33% of the clusters formed in the graph. The authors used the Intact-Group Comparison Method consisting of 7 research steps: Design Selection, Determination of Representative Sample, Instrumentation, Implementation of Experiment, Data Collection and Analysis, Data Interpretation and Analysis, and Experiment Conclusions.With the result of clusters increased from 3 to 4 clusters. This research can be developed using True Experiment Design and the addition of Sentiment Analysis.

Copyrights © 2018