In this work, we propose an algorithm based on transitive extension for finding friend groups in blogosphere to perform social network analysis. In today's blog service environment, the establishments of friend relationships are always unidirectional, and the recognition of both ends is not necessary. Therefore, general methods will find either too small or too few cliques from friend groups. This isbecause the bidirectional link is built incompletely in the social network under such circumstances. To solve this problem, we propose the 1.5-clique extension method to derive better social network structures for finding friend groups, and we use Wretch (www.wretch.cc) as a case for analysis. We further make a comparison among the results of finding groups in the original social network, and its 1-clique extension, 1.5-clique extension, and 2-cliqueextension. The experimental results suggest that our proposed method is effective and promising.