<html><head><meta http-equiv="Content-Type" content="text/html; charset=us-ascii"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class=""><blockquote type="cite" class=""><blockquote type="cite" class="">...Finding cliques is NP-complete. This very strongly suggests that the problem you suggest is as well.<br class=""></blockquote></blockquote><div><blockquote type="cite" class=""><div class=""><div class=""><br class="">Most NP complete problems have approximate or near optimal solutions that are merely polynomial...</div></div></blockquote><div>However, some NP-complete problems are not approximable - a celebrated result a number of years back. Unfortunately, the clique problem is one of them. (<a href="http://www.cs.umd.edu/~srin/PS/clique-jou.ps" class="">http://www.cs.umd.edu/~srin/PS/clique-jou.ps</a> has a discussion if you really want to get into the details).</div><br class=""><blockquote type="cite" class=""><div class=""><div class=""><blockquote type="cite" class="">There are going to be a lot of naturally-occurring, legitimate (closed) groups of nodes in a network of buyers....<br class=""></blockquote><br class="">Legitimate fans are likely to be fans of other musicians as well, and likely to have genuine interactions with fans of other musicians.<br class=""></div></div></blockquote></div>Only experiment will determine if this is really true, and to what degree - but my guess is you'll have plenty of false positives. In the end this will be a ROC problem - you have to trade off among sensitivity, specificity, and so on. Since we're talking about an active opponent here, not random noise ... expect this to be very hard.<div class=""><div class=""> -- Jerry</div></div><div class=""><br class=""></div></body></html>