Propuesta de un modelo para análisis de prevención de lavado de dinero en Ecuador basado en Machine Learning

Abstract

The general objective is to develop a model in the prevention of money laundering to identify patterns and improve decision-making based on Machine Learning technology. The methodology for the extraction and analysis of scientific articles, digital libraries are explored, and the Random Forest algorithm is adopted according to the extraction and analysis of scientific articles that recommend the most appropriate algorithm. A total of 22 articles were obtained for analysis, and the following anti-money laundering items were found: The largest sector that uses ML is Banking, the first ML algorithm they use is Random Forest, the largest entities that use ML are Commercial, the first activity that is carried out is Fraud, the first technique that is performed is Classification, the largest tool used with ML algorithms is Deep Learning. A model based on the Random Forest classifier is proposed because it is based on probability scoring, selection of the strangest vectors, and selection of the least bizarre aggregate transactions. Of course, there is a lot of room for improvement in the proposed model. In addition, you need a very large dataset with deep fields or attributes.