Accounting for Social, Spatial, and Textual Interconnections

摘要: This is a chapter about what link analysis and data mining can do for criminal investigation. It long complex chapter, in which variety of techniques topics are accommodated. divided two parts, one methods, the other real-case studies. We begin by discussing social networks their visualisation, as well unites them with or distinguishes from (which itself historically arose disciplinary context ergonomics). Having considered applications to investigation, we turn crime risk assessment, geographic information systems mapping crimes, detection, then multiagent architectures application policing. challenge handling disparate mass data, introduce reader warehousing, XML, ontologies, legal financial fraud ontology. A section automated summarisation its law followed discussion text law, on support vector machines retrieval, classification, matching. follows, stylometrics, determining authorship, handwriting identification automation, questioned documents evidence. next discuss clustering, series analysis, association knowledge discovery databases; then, inconsistent data; rule induction (including law); using neural context; fuzzy logic; genetic algorithms. Before turning case studies mining, take broad view digital resources uncovering perpetration: email computer forensics, intrusion detection. consider Enron database; coalitions SIGHTS system, recursive mining. steganography, detection (the use learning techniques, masquerading, honeypots trapping intruders). Case include, example: investigating Internet auction NetProbe; graph malware Polonium; Coplink; project U.S. Federal Defense Financial Accounting Service; extraction tools integration tool; Poznan ontology model fuel fraud; fiscal Pisa SNIPER project.