Privacy　is　a　fundamental　issue　in　big　data.　Meanwhile,　determining　semantic　relationships　between　words　and　phrases　in　privacy　is　required　for　effective　privacy　protection　to　the　data　that　originates　from　a　variety　of　sources,　a　main　characteristic　of　big　data.　WordNet　has　been　used　as　one　of　the　most　popular　ways　of　measuring　semantic　similarity　between　words.　In　this　paper,　through　comparison　analysis,　we　show　that　WordNet　is　not　very　adequate　for　measuring　semantic　similarity　or　relatedness　between　words　when　concerning　privacy.　The　analysis　consists　of　an　experiment　to　get　human　rating　scores　as　the　benchmark　dataset　and　the　comparison　between　results　from　WordNet　based　measures　and　the　benchmark　dataset　to　reach　the　conclusion.