Not logged in.
Quick Search - Contribution
Contribution Details
| Type | Bachelor's Thesis |
| Scope | Discipline-based scholarship |
| Title | Location-based Open Source Intelligence to Infer Information in LoRa Networks |
| Organization Unit | |
| Authors |
|
| Supervisors |
|
| Language |
|
| Institution | University of Zurich |
| Faculty | Faculty of Business, Economics and Informatics |
| Date | 2023 |
| Abstract Text | This thesis introduces and evaluates a novel platform that uses Open-source intelligence (OSINT) to identify a primary subject and an associated event using publicly accessible data. As a starting point, the platform utilizes LoRa (Long Range) datasets. This novel tool will make use of web scraping techniques, the power of OpenAI's large language model GPT-3.5, and a custom matching score algorithm. The objective is to collect a comprehensive image of the primary subject and infer potential participants of the specific location and time covered by the LoRa dataset. Evaluating our approach demonstrates its effectiveness in identifying 14 out of 16 actual participants, showcasing its ability to create a relevant dataset of potential participants. Looking at the accuracy, the model manages to achieve a precision score of 0.75, while the recall score of 0.46 indicates some true positives were not captured. The results reflect the difficulty in identifying participants in a private event with a limited public presence. Despite the challenging scenario, this tool represents an innovative approach to merging OSINT techniques with LoRa data. Future work will focus on enhancing the tool's robustness, expanding its coverage to additional social media platforms, improving adaptability across diverse scenarios, and exploring advanced language models. |
| PDF File |
Download
|
| Export |
BibTeX
|