巨量資料的獲取使得政府在打擊犯罪方面更有效率，然而目前仍然面臨相關資源不足的問題。由於犯罪數據資源分佈獲取不均，政府部門難以獲取充足有效的犯罪活動資訊。這方面數據的匱乏限制了犯罪模式分析工作的進展。在一些開發中國家，由於缺乏詳細的街道地圖資料，使得現實中的諸多地址未能在地圖上座標化，導致現有的地圖資料在地理資訊分析上無法物盡其用。但通過結合多維度的犯罪記錄資訊可以發現某些犯罪事件在地理上具有相似性，進而得到相似犯罪事件附近的座標並將此犯罪事件用於進一步的聚類分析，從而得以解決上述問題。爲了提高犯罪事件之間的關聯性，本文首先使用模糊聚類方法降低原始資料的多樣性。本論文提出了一種結合空間、時間與多種犯罪資料維度的方法以解決無座標的犯罪事件定位問題，並協助犯罪事件的聚類分析。

The collection and storage of mass amounts of data have made crime fighting more efficient and effective. However, a common problem encountered by law enforcement is insufficient resources. Coupled with a lack of proper information on criminal activities because of data flaws can contribute to the improper use of the resources. Deficient data can cause limitations in discovering useful patterns. In developing countries, a common data issue can arise when the scope of public maps for streets is inadequate, consequently those data become geographically worthless. The deficiency of scope of public maps causes many post addresses to be rendered ungeocodable. However, this problem can be addressed by associating dimensions within the crime records to discover which crimes are geographically similar and obtain a nearby coordinate which will allow the event to be used in clustering. To improve association between events, fuzzy clustering applied to the raw data first can reduce variety among the data. In this thesis, the relationship between the spatial and temporal components, and crime dimensions are associated to place the ungeocodable crime events on the map and aid crime clustering.

摘要. . . . . . . . . . . . . . . . . . . . . . . . . . i
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . ii
Acknowledgement. . . . . . . . . . . . . . . . . . . . . . . . . . iii
List of Tables. . . . . . . . . . . . . . . . . . . . . . . . . . vi
List of Figures. . . . . . . . . . . . . . . . . . . . . . . . . . vii
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3 Method. . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.2 Fuzzy Clustering to repair Text Data . . . . . . . . . . . . . 13
3.3 Data Reduction . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.4 Association . . . . . . . . . . . . . . . . . . . . . . . . 16
3.4.1 First Stage Clustering . . . . . . . . . . . . . . . . . . . . 16
3.4.2 Second Stage Clustering - Association . . . . . . . . . . . . . . . . 18
3.4.3 Minor Cluster Aggregation . . . . . . . . . . . . . . . . . . . . . . 19
4 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 31
References. . . . . . . . . . . . . . . . . . . . . . . . . . 33

[1] Satya Chakravorty. Identifying crime clusters: The spatial principles.MiddleStates Geographer, 28:53–58, 1995.
[2] Lawrence W. Sherman. Hot spots of crime and criminal careers of places.Crimeand Place, 4:35–52, 1995.
[3] IBM. Ibm big data analytics technology helps durham police reduce crime, 2013.
[4] Tony Grubesic and Alan Murray. Detecting hot spots using cluster analysis andgis. InProceedings of the Fifth Annual International Crime Mapping ResearchConference, page 26, 2001.
[5] Jyoti Agarwal, Renuka Nagpal, and Rajni Sehgal. Crime analysis using k-meansclustering.International Journal of Computer Applications IJCA, 83(4):1–4,2013.
[6] Lawrence Sherman and David Weisburd. General deterrent effects of police patrolin crime hot spots: A randomized, controlled trial.Justice Quarterly, 12(4):625–648, 1995.
[7] John Eck, Spencer Chainey, James Cameron, and R Wilson. Mapping crime:Understanding hotspots. pages 1–79, 2005.
[8] E Limoges. Improvement of decennial census small-area employment data: Newmethods to allocate ungeocodable workers. InTransportation Research BoardConference Proceedings, volume 2, 1997.
[9] Jieming Shi, Nikos Mamoulis, Dingming Wu, and David W Cheung. Density-based place clustering in geo-social networks. InProceedings of the 2014 ACMSIGMOD international conference on Management of data, pages 99–110. ACM,2014.
[10] Shaoxu Song, Chunping Li, and Xiaoquan Zhang. Turn waste into wealth: Onsimultaneous clustering and cleaning over dirty data. InProceedings of the 21thACM SIGKDD International Conference on Knowledge Discovery and Data Min-ing, pages 1115–1124. ACM, 2015.23
[11] Shyam Varan Nath. Crime pattern detection using data mining. InWeb Intel-ligence and Intelligent Agent Technology Workshops, 2006. WI-IAT 2006 Work-shops. 2006 IEEE/WIC/ACM International Conference on, pages 41–44. IEEE,2006.
[12] Peter Rogerson and Y Sun. Spatial monitoring of geographic patterns: an applica-tion to crime analysis.Computers, Environment and Urban Systems, 25(6):539–556, 2001.
[13] Tan Pang-Ning, Michael Steinbach, Vipin Kumar, et al. Introduction to data min-ing. InLibrary of congress, volume 74, 2006.
[14] Ralph B Taylor.Breaking away from broken windows: Baltimore neighborhoodsand the nationwide fight against crime, grime, fear, and decline. Westview PressBoulder, CO, 2001.[15] Orlando J P ́erez. Gang violence and insecurity in contemporary central america.Bulletin of Latin American Research, 32(s1):217–234, 2013.