監視器錄影系統與竊盜犯罪率之時空分析
- 發布日期：
- 最後更新日期：109-05-13
- 資料點閱次數：285
資料來源:http://ndltd.ncl.edu.tw/cgi-bin/gs32/gsweb.cgi/ccd=MtgIBv/record?r1=14&h1=5
:::
監視錄影系統與現今社會治安息息相關。又因竊盜犯罪發生率高且影響民眾治安觀感甚深，故本研究選擇臺北市2008年至2014年間之三種竊盜犯罪案件(普通及住宅竊盜、機車竊盜、汽車竊盜)為研究對象。分別從時間、空間角度來分析對竊盜犯罪之影響。時間方面，以年、月和日等尺度；空間方面，透過地理資訊系統之視覺化及空間運算功能，來探討裝設監視錄影系統影響竊盜犯罪犯罪熱區分布情形，及犯罪數增減之趨勢。除了監視錄影系統密度之外，本研究亦選取與竊盜犯罪相關之人文社經變項，進行空間迴歸分析，深入分析竊盜犯罪相關因素。本研究使用次級資料，資料來源為國土資訊系統社會經濟資料庫、臺北市政府警察局委託中央警察大學進行臺北市錄影監視系統運用於犯罪偵防效能之研究，分析軟體為SPSS、Arc GIS、R與Geo Da。
研究結果顯示：(1)時間方面，竊盜犯罪隨著年份增加而遞減、不同竊盜種類分別有不同高峰發生之月份與時段；監視器設置後機車、汽車竊盜犯罪數減少程度達統計顯著，而普通及住宅竊盜犯罪數減少程度未達統計顯著。(2)空間方面，普通及住宅竊盜與機車竊盜犯罪有相似的聚集分布並集中於鬧區，而汽車竊盜發生在偏遠地區並遠離市中心。(3)空間迴歸分析部分，空間迴歸分析以地理加權迴歸分析具有最佳之配適度。監視器密度、人口密度、高教育程度之人口比、離婚率、遷入出人數與收入對於竊盜犯罪有顯著影響。唯監視器密度對汽車竊盜影響並不顯著。
本研究建議方面，發掘潛在之影響因素，可將民間監視器之數據納入後續研究，更貼近監視器真實功效；另可透過筆錄內容確定犯罪發生時間，以增加研究準確度；考量臺北市與新北市鄰近之地區特性，可減少邊際效應對研究成果之影響。Closed circuit television system (CCTV) plays an important role in crime prevention and control in a city. The aim of this study was to estimate the effectiveness of CCTV in Taipei city. Exactly 76,750 theft crimes, occurred between 2008 and 2014, were chosen as the study data; and are separated into 3 subgroups: burglary, motorcycle theft, and vehicle theft. The main analysis used in this research includes three parts: temporal analysis, spatial analysis, and regression modelling. For temporal analysis, this study used three different timescales: yearly, monthly, and hourly, to check the rush time of crimes. For spatial analysis, ANN and Kernel estimation were used to define the cluster patterns and hot zones of the crime data. Additionally, four spatial regression models were used to estimate the relationship between crime rate and its possible related factors, such as CCTV density, population density, and devoice rate. The crime data was obtained from Taipei City police department, and the external factors, such as demographic and socio economic variables, were downloaded from Taiwan Geospatial One-Stop Portal (TGOS) websites.
The results show that (1) the number of crime occurrences was significantly reduced after CCTV installation in 2010 and the time distributions of three crime types are all different. (2) For spatial analysis, both hot zones of burglary and motorcycle theft were concentrated in the urban area, while that of vehicle theft was located in the rural area; yet all three subgroups were defined as cluster data. Moreover, among the applied regression models, geographical weighted regression (GWR) model performed best, then spatial error model, then spatial lag model, and the least applicable model was OLS model. Furthermore, the significant variables were identified: population density, population ratio of higher education, divorce rate, immigration and emigration, and average income; whereas CCTV density had an insignificant effect on theft crime.
Although the results prove the importance of this research, there were two main limitations which need to be addressed in future studies: (1) the designed scope of the study only included public CCTV and neglected private CCTV which provided a partial evaluation of the effectiveness of CCTV. (2) Inconsideration of edge effects between Taipei city and New Taipei city boundary simplified our data, however there are strong interactions and a large amount of commuters between these two cities which need to be taken into account.