臺中市監視錄影系統(以下,簡稱CCTV),自2005年建置迄令已屆滿10年,但在犯罪預防成效上,仍面臨以下諸多困境問題,例如:(1)近10年來市區經濟發展和環境變遷,CCTV建置點位是否符合區域比例之原則?(2)是否將原建置CCTV點位進行全面性之量化檢測評估,以利機動調整撙節開支?(3)在民眾普遍認為CCTV建置率不足的情況下,如何運用最新數位影像科技,優先建置警察專屬GIS地理資訊系統,以提升CCTV即時監控之效能。故本研究係結合SAS研究團隊,運用大數據(big data)探勘技術,嘗試以SAS Overview「進階分析犯罪預測解決方案」套裝軟體,以SAS Visual Analytic視覺化統計測量作為使用者操作介面,採用時間序列預測方式,根據過去刑案的發生破獲件數,預測未來偵防犯罪之策略作為。以2011-2014年臺中市全般刑案發生破獲數、調閱CCTV破案數(含交通事故)及建置CCTV點位數等原始數據資料,進行關聯性分析與資訊視覺化分析,期能建立「視覺化管理」模式,成為端末使用者簡易方便之操作介面。並藉由電子地圖自動選擇最佳化方式,使其與GIS地理資訊系統接軌,以達成CCTV機動靈活監控之願景。
The Taichung City close d-circuit television (CCTV) system has been in commission since 2005; however, there are still several questions regarding its effectiveness in crime prevention These include: (1) Do the CCTV camera locations conform to the local scale after a decade of economic and environmental changes? (2) Should the CCTV camera locations undergo quantitative inspection and evaluation to adjust and cut back expenditures? (3) As the public feels that CCTV cove rage is insufficient, how can the latest digital imaging technologies be used to establish a police-dedicated geographic information system (G IS) in order to improve the efficacy of real-time monitoring. As such, this study integrated a SAS team and big data mining technologies in an attempt to use a SAS Overview advanced analytic crime prediction soft ware package and a SAS Visual Analytics user interface in order to determine future crime prevention strategies based on time series predict ion. Raw data for 2011-2014, including the number of offenses cleared in Taichung City, the number of cases solved consulting CCTV (including traffic accidents), and the number of CCTV camera locations, was use d to conduct correlation analysis and data visualization analysis in order to establish a visualization management model. This model was to be a simple end user interface integrated with the GIS via electronic map automatic selection optimization in order to create a flexible e CTY-driven monitoring grid.