Using Data Mining and Machine Learning to Detect Cryptocurrency Risk
An overview of how natural language processing methods (NLP) can be used to tag adverse news reports for cryptocurrency monitoring - a joint project with NUS Business Analytics students.
January 12, 2021 · 1 min read
Cylynx partners with educational institutions to drive research in machine learning on fraud detection. In one of our recent projects, we partnered with a group of National University of Singapore (NUS) Business Analytics students to examine how natural language processing (NLP) could be used to tag adverse news reports. Such a solution can assist in automatic monitoring of cryptocurrency entities and flag cyber breaches on suspected fraud incidents when it is mentioned in the media.
The students consisting of Lai Yan Jean, Lee Jing Xuan, Valary Lim Wan Qian and Xu Pengtai developed a pipeline to scrape conventional news and social media, before running it through an NLP model (RoBERTa) to determine its risk score. Check out their Medium post article for a more detailed write-up of the approach and solution.
As attention in the virtual asset space increases, investors and regulators are paying more scrutiny on the compliance practices of crytocurrency firms. Social media and news monitoring is a good way to keep up to date with the activities on the ground. An automated risk scoring solution helps eliminate the tedious work of crawling through news content by surfacing only the relevant ones that is of concern.
Are you looking to implement automated social media monitoring of organizations or searching for an out of the box solution to monitor virtual asset service providers? Contact us for more information about our adverse news screening services.