Abstract
Studying Online Social Networks such as Facebook, MySpace, Twitter…etc, will help us understand the formation of relationships between people in social networks. Nowadays, the most popular Online Social Network Sites is Facebook. There have been more than 600 million accounts registered in Facebook. However, few researches has used real-world data in their researches due to difficulties of having these data. Assume there are 3 nodes A, B, and C in a network graph. Node A is connected to node B and node C, and node B and C are not connected. In this case, Node A is the mutual friend of node B and C, and, more importantly, we focus on how node A affects the probability of establishing a link between B and C. In our definitions, Node B and C will be connected with probability a, and will not be connected with probability 1-a. Our goal is to find out the a value in real-world data. To acquire these real-world dataset, we have some dataset which are collected by former researchers. Though we calculate the value a and get our results, we are still not able to study the link establishment and link vanishing properties in these dataset, which are provided by other researchers. For this reason, we develop a program to collect information from Online Social Networks. This program has the following features: 1. Collect social graphs in our ways. 2. Observe the changing of network graphs in a period of time. 3. Use collected data to test our new algorithms. We name this program as Web Crawler, which “crawls” web pages to collect user’s information in OSNs. For observing the ever-changing networks with link establishing and vanishing, we modified the program to collect user information in a period of time and compare it with former collected data.