In　the　space　clustering　algorithm,　because　of　the　choice　of　the　value　of　k　and　the　problem　of　the　non-clear　＇elbow　point＇　of　the　elbow　method,　this　paper　introduces　logarithmic　function　and　determines　the　initial　clustering　center　on　the　basis　of　the　properties　of　exponential　function,　weight　adjustment,　bigotry　term　and　the　basic　idea　of　elbow　method,　and　proposes　an　improved　k-value　selection　algorithm.　Combined　with　the　fully　adaptive　spectral　clustering　algorithm,　the　global　terrorist　attack　data　are　clustered.　It　effectively　solves　the　problem　that　the　selection　of　k　value　is　not　clear　and　the　outlier　can　not　be　separated　in　the　clustering　process.　The　experimental　results　show　that　the　clustering　method　proposed　in　this　paper　can　not　only　determine　the　k　value　quickly　and　accurately,　but　also　achieve　better　clustering　effect.　CesiumJS,　WebVR　and　speech　recognition　technology　are　used　to　visually　display　and　interact　the　results.　©　2019　Association　for　Computing　Machinery.