Abstract
The main object of this paper is to present a network formation model for social network. To start with, we create a fully connected graph that has m0 vertices. At every time step, one vertex v is introduced into the network. The new vertex v randomly attaches to one existing vertex with equal probability. Then, each neighbor of the attached vertex forms a connection to the new vertex v with probability a. We call this operation triad formation. We derive the mean degree, degree distribution and the clustering coefficient for this model. We show that the clustering coefficient is non-vanishing as the network grows. The expected degree and expected number of triangles are tunable simply by changing the control parameter a.