Represents to Complex Networks and Prediction of VDO Online with Data Mining

Abstract

In order to work through the internet, there are a variety of service delivery models to communicate the information to each other by using social media to reach people in all levels and leading to social networking such as sharing photos and watching video online through the Facebook, Twitter and YouTube which are popular nowadays. Therefore, in this paper, the users information from social media related to online video with the aim to 1) represent to complex network using Pajek program, 2) classify social media using SOM technique, and 3) find out a decision tree techniques to classify the high accuracy and using decision tree to generate a rule of decision applied to the niche market. The data set from website of pewinternet.org consisted of 1,003 people. The instruments of the research were  Weka , Orange Canvas and Pajek.  The findings found that the relation between social media users presents complex networks in the form of a "small world". A cluster of users with high level education was employed full time and watching video online through social media the most. The decision trees with ID3 technique contributed to the highest accuracy at 98.7179 % and 12 rule base of ID3 technique.