A Novel Solution For Anti-Money Laundering System

Abstract

In the age of unpredictable fluctuations of technology, disorganized detection has been recently figured out in most of present-day anti-money laundering systems. These obstacles are attributed to certain reasons associated with applying handcrafted manipulation in the long list of principles and having the shortage of real datasets about banking purchasers or the customers’ information. This article demonstrates such an innovative approach to evaluate the data in terms of suspicious behaviors, clients’ relationships, the awareness for the customers retrieval from the financial sector in social media platforms. The applicable datasets consisting of above 20000 sample records on Kaggle is the main resource for our service. Each entry was compiled from content of collected documents and was attached to the descriptions measuring positivity or negativity in catching money laundering. They were used to qualify the model in AutoML supplied by Google Cloud Artificial Intelligence. After having been satisfied the sentiment standard with a performance accuracy approximately 0.85, we attempted to forecast the sentimental design for all searched outcomes connected with the clients to distinguish badly known companies. The output is a beneficial tool for the companies getting used to realizing unauthorized clients. In other words, instead of having no information about new clients in Know Your Customer of anti-money laundering inspections, it is more helpful to utilize this service without wasting too much time and money for a huge number of other sites out there.