With　the　change　in　people＇s　lifestyle　and　travel　mode,　understanding　the　individual　and　population　mobility　patterns　in　urban　areas　remains　to　an　outstanding　problem.　Pervasive　mobile　communication　technologies　generate　voluminous　data　related　to　human　mobility,　such　as　mobile　phone　data.　To　further　study　the　characteristics　of　returning　and　exploration　patterns　of　human　movement　in　urban　space,　a　multi-index　model　is　proposed　based　on　the　original　radius　of　the　gyration　index.　In　this　paper,　the　classification　mechanism　of　a　single　ratio　of　the　radius　of　gyration　for　k-explorers　and　k-returners　is　illustrated.　Some　disadvantages　of　this　mechanism　are　noted.　A　few　indices　of　the　model　are　proposed　for　deep　mining　of　data　on　human　mobility　exploration　and　returning　characteristics.　Taking　a　mobile　phone　data　during　an　entire　month　as　a　sample,　and　after　data　processing　on　the　Spark　platform,　the　characteristics　of　various　indicators　and　their　correlations　are　analyzed.　The　classification　effects　of　different　spatial　indices　for　human　exploration　and　returning　are　compared　by　using　a　support　vector　machine　and　the　binary　classification　algorithm　and　are　further　compared　with　existing　research　results.　The　differences　in　the　classification　effects　of　these　indicators　are　analyzed,　which　is　helpful　for　in-depth　studies　of　urban　mobility　patterns.