Thispaperaddressesanimportantandinterestingproblem-automaticallyidentifyingadultaccountsonSinaWeibo.Theauthorsproposetwosetsofbehaviorindicatorsforadultgroupsandaccounts,andfindthatadultgroupsandaccountshavedifferentbehavioraldistributionswithnon-adultgroupsandaccounts.Thenanovelrelation-basedmodel,whichconsiderstheinter-relationshipsamonggroups,individualaccountsandmessagesources,isappliedtoidentifyadultaccounts.Theexperimentalresultsshowthatcomparedwithstate-of-the-artmethods,theproposedmethodcanimprovetheperformanceofadultaccountidentificationonSinaWeibo.Overall,thearticleiswellorganizedanditspresentationisgood.However,someminorissuesstillneedtobeimproved:(1)TheauthorsshouldsummarizethemaincontributionsofthispaperinSection1.(2)InSection4.2,theauthorsmentionedthat“AgroupwillattainavalueveryclosetoonGACSifallitsaccountshaveentirelycopiedtheirowntexts,imagesorcontactinformation”.However,accordingtoEquation8,contactinformationisnotconsideredwhencomputingGACS.(3)InAlgorithm1onPg.17,itseemsthat“t=t+1”shouldbeaddedafterline6.(4)IsuggestthatthelimitationofthisworkshouldbediscussedinSection9.(5)Thereareafewtyposandgrammarerrorsinthispaper.