Conceptual graph is a graph in which nodes are concepts and the edges indicate the relationship between them. Creation of conceptual graphs is a hot topic in the area of knowledge discovery. Natural Language Processing (NLP) based conceptual graph creation is one of the efficient but costly methods in the field of information extraction. Compared to NLP based methods, Statistical methods have two advantages, namely, they are language independent and more computationally efficient. In this paper we present an efficient statistical method for creating a conceptual graph from a large document collection. The documents which are used in this paper are from Wikipedia collection because of their rich and valid content. Moreover, we use the final conceptual graph to suggest a list of similar keywords for each unique concept or combination of concepts. Also, we will show the viability of our approach by comparing its result to a similar system called the Wordy system.