Over the past two decades, technology has developed tremendously, at an almost exponential rate. While this development has served the nation in numerous different positive ways, negatives have also emerged. One such negative is that of computer crime. This criminality has even grown so fast as to leave current digital forensic tools lagging behind in terms of development, and capabilities to manage such increasing and sophisticated types of crime. In essence the time taken to analyse a case is huge and increasing, and cases are not fully or properly investigated. This results in an ever-increasing number of pending and unsolved cases pertaining to computer crime. Digital forensics has become an essential tool in the fight against computer crime, providing both procedures and tools for the acquisition, examination and analysis of digital evidence. However, the use of technology is expanding at an ever-increasing rate, with the number of devices a single user might engage with increasing from a single device to 3 or more, the data capacity of those devices reaching far into the Terabytes, and the nature of the underlying technology evolving (for example, the use of cloud services). This results in an incredible challenge for forensic examiners to process and analyse cases in an efficient and effective manner. This thesis focuses upon the examination and analysis phases of the investigative process and considers whether automation of the process is possible. The investigation begins with researching the current state of the art, and illustrates a wide range of challenges that are facing the digital forensics investigators when analysing a case. Supported by a survey of forensic researchers and practitioners, key challenges were identified and prioritised. It was found that 95% of participants believed that the number of forensic investigations would increase in the coming times, with 75% of participants believing that the time consumed in such cases would increase. With regards to the digital forensic sophistication, 95% of the participants expected a rise in the complexity level and sophistication of digital forensics. To this end, an automated intelligent system that could be used to reduce the investigator’s time and cognitive load was found to be a promising solution. A series of experiments are devised around the use of Self-Organising Maps (SOMs) – a technique well known for unsupervised clustering of objects. The analysis is performed on a range of file system and application-level objects (e.g. email, internet activity) across four forensic cases. Experiment evaluations revealed SOMs are able to successfully cluster forensic artefacts from the remaining files. Having established SOMs are capable of clustering wanted artefacts from the case, a novel algorithm referred to as the Automated Evidence Profiler (AEP), is proposed to encapsulate the process and provide further refinement of the artefact identification process. The algorithm led to achieving identification rates in examined cases of 100% in two cases and 94% in a third. A novel architecture is proposed to support the algorithm in an operational capacity – considering standard forensic techniques such as hashing for known files, file signature analysis, application-level analysis. This provides a mechanism that is capable of utilising the A E P with several other components that are able to filter, prioritise and visualise artefacts of interest to investigator. The approach, known as Automated Forensic Examiner (AFE), is capable of identifying potential evidence in a more efficient and effective manner. The approach was evaluated by a number of experts in the field, and it was unanimously agreed that the chosen research problem was one with great validity. Further to this, the experts all showed support for the Automated Forensic Examiner based on the results of cases analysed.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:701350 |
| Date | January 2016 |
| Creators | Al Fahdi, Mahmood |
| Publisher | University of Plymouth |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://hdl.handle.net/10026.1/8090 |
Page generated in 0.0017 seconds