Social networks and virtual communities has become a popular communication tool among Internet users. Millions of users share publications about different aspects: educational, personal, cultural, etc. Therefore these social sites are rich sources of information about who can help us solve any problems. In this paper, we focus on using the written comments to recommend a person who can answer a request. An automatic analysis of information using text mining techniques was proposed to select the most suitable users. Experimental evaluations show that the proposed techniques are efficient and perform better than a standard search.