留学生计算机科学与技术专业硕士课程论文范文1 Introduction
The report summarizes the work I did in the last few months.I do some study for the background of this research which contains the theories, algorithm that detect and analyse the evolving communities, and so on. In order to do some research on the evolving communities, some related theories about communities, Clique percolation method ( CPM ) should be studied firstly. Then chapter three emphasize the frame of community detecting. The next chapter explain the method of analysing the evolving communities. In the last part of this report I show the previous work in the related area.
2 Related theories
This chapter gives a brief description of related theories about community detecting, which contain the introduction ,the formal expression, the types and the topological properties of network, the concept and properties of community, and the principle of CPM.
2.1 Network
2.11 Basic concept
In the real life, some systems that involve multi-individual and some of them interact with others can be abstracted as network where each individual is a node of network and edge of network represent the relationship or interaction of individuals. In the co-authorship network the nodes in are authors, which are linked if they co-author the paper. The network can be generated as graph G which can be illustrated as {V, E}, based on the the properties of V and E, network can be divided into several styles. If a random pair of nodes (i,j) and (j,i) has the same edge, then we call this network is undirected network, and the co-authorship network belongs to undirected network[1]. In the weighted network every edge can be marked by some property, for example, types and weight, and the the size of the weight present the degree of association. However, in co-authorship network the number of the weight stands for the times of collaboration.
2.1.2 Related property
Plenty of studies found that, related network such as co-authorship network is greatly difference of random network.
1 small-world effect
Even though the co-authorship network is very huge, the average path length of network which is defined as the average distance between any two nodes in network is real small.
presents the distance between the nodes i and j
The small-word is firstly proposed by Hungarian--F.Karinthy in 1929, he thought any two persons can be linked by average six lines that are composed by six persons[2]. After that, in 1967, Milgram who are the social psychologist of Harvard University proposed the six degrees of separation that confirm the the property of small-word[3]. Hereafter, people use other method to test the six degrees of separation hypothesis. In , professor Grossman in university of auckland proposed the Erdos number project tested the small-word in collaboration network of mathematician[5], and Elmacioglu do some research on savant collaboration network based on the DBLP data set, and test the six degrees of separation phenomenon[4]. |
