| Abstract |
In the modern era, mobile phones have become universal devices capable of serving various essential purposes, including real time crime scene capture, reporting, and rapid police alerting. This thesis introduces different innovative approaches for detecting suspicious activities using mobile sensor data, with a focus on post-crime detection and criminal tracking through the analysis of sensor data patterns. The research encompasses the development of a robust digital forensic system and investigation techniques that harness the potential of mobile sensor data. The thesis outlines several unique approaches to pattern recognition and crime investigation. The first approach involves the creation of a modified sub-space K-NN (msK) algorithm, designed to efficiently recognize sensor data patterns indicative of suspicious activities. This msK approach was also compared with traditional Gaussian SVM. To aid this research, an open-source mobile application, the "Evidence Collector (EC) app", was developed to enable t
In the modern era, mobile phones have become universal devices capable of serving various essential purposes, including real time crime scene capture, reporting, and rapid police alerting. This thesis introduces different innovative approaches for detecting suspicious activities using mobile sensor data, with a focus on post-crime detection and criminal tracking through the analysis of sensor data patterns. The research encompasses the development of a robust digital forensic system and investigation
techniques that harness the potential of mobile sensor data. The thesis outlines several unique approaches to pattern recognition and crime investigation. The first approach involves the creation of a modified sub-space K-NN (msK) algorithm, designed to efficiently recognize sensor data patterns indicative of suspicious activities. This msK approach was also compared with traditional Gaussian SVM. To aid this research, an open-source mobile application, the "Evidence Collector (EC) app", was developed to enable t
|
