DEPT. OF DATA ANALYSIS, DECISION MAKING AND FINA
Banks in Russia have experienced a rise in client complaints regarding fraudulent monetary transfers made under the influence of social engineering. In collaboration with students at the Financial University in Moscow, the team used artificial intelligence to identify a speaker's emotional state using acoustic markers. Three models were developed and tested with eleven datasets to enhance accurate identification. Based on the results, the authors provide recommendations on the most salient vocal features and classification models to aid with fraud detection.
Worcester Polytechnic Institute
Interactive Qualifying Project
All authors have granted to WPI a nonexclusive royalty-free license to distribute copies of the work, subject to other agreements. Copyright is held by the author or authors, with all rights reserved, unless otherwise noted.
Restricted-WPI community only
Humanities and Arts
Available for download on Wednesday, May 12, 2021
Your accessibility may vary due to other restrictions.