Implementation of FaDO in React.js : Implementation of FaDO in React.js
Abstract: This report describes the implementation of FaDO as a REACT.JS web-based tool. FaDO is a machine learning method for real-time fraud detection: it flags (financial)transactions as either fraudulent or normal in a real-time fashion. FaDO implementssemi-supervised machine learning for this task: in semi-supervised learning, only a small portion of the presented data needs to be labeled. This tool is designed as a filter that works together both with existing automatic fraud filters and human experts. The contribution of this work is a dynamical webpage that implements and demonstrates the approach. Herefore, the REACT framework and the JSX programming language were used. The emphasis of this tool is put on real-time in-browser processing of all data. In order to illustrate the tool, a realistic artificial transaction generator was implemented. The tool then allows to organise different fraud filters to select fraudulent transactions from this high-frequency stream. This report gives also preliminary key performance indicators (KPIs) of the approach. A comparison to a realistic hand-crafted expert system is given.
AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)