Money Laundering Detection System with Intelligent Agents

Over the years, people acquire money illegally from public treasury and launder it by depositing it as clean money into the banking system. Hence, the laundered money is integrated back into the financial system concealing the illicit sources. Money laundering is a major challenge and a threat to both financial institutions and government. This research is aimed at developing an Enhanced Multi-agent Money Laundering Detection System for monitoring cash flows in individual and corporate accounts in deposit money banks and other financial institutions, so as to detect illicit fund movement. In an effort to checkmate money laundering, threshold is used by financial institutions to check the volume of money an individual or company can transact in a single transaction. This is defective as the transaction can be broken into pieces to avoid being detected. The existing Money Laundering detection systems lack intelligence on detecting money laundering when it evades the stated threshold. This work aims at developing an Enhanced Multi-Agent Money Laundering Detection System with time frequency analysis. The system was designed to monitor financial institutions monetary transactions to detect money laundering activities. The intelligent agent system was designed to perform some tasks autonomously, which means that the system can act independently of humans. A set of autonomous type of behaviours for the agent class, including reactive, proactive, and cooperative behaviour was designed. Both autonomous and semi-autonomous agents rely on data (customer daily transaction history) which is the agent’s perception or awareness of its environment. Various kinds of intelligence are supported by this kind of data. Business rules form an important part of the knowledge base for software agents to perform delegated tasks. Also the system uses time frequency analysis to detect money laundering which involves multiple account or multiple transactions. Object-Oriented Analysis and Design Methodology (OOADM) was adopted in this research work. The software was developed using Hypertext Markup Language (HTML), Hypertext Preprocessor (PHP), My Structured Query Language (MySQL), Cascaded Style Sheet (CSS), Java Script, Dream weaver, and Fireworks. The software performance was tested using accuracy in money laundering detection as the key performance index (KPI). In the software test carried out 97% accuracy in detecting money laundering was achieved. Keywords: Bank verification Number (BVN), Cyber security, Cyber crime, Money laundering, Intelligent agents, Time-frequency analysis