Breaking the chain of Illicit Finance

Utilising network CLIQUE detection for advanced anti-money laundering.

Before filling out the registration form, please read the Privacy notice pursuant to Article 13 of EU Regulation 2016/679

Invalid Input
Invalid Input
Invalid Input
Invalid Input
Invalid Input
Invalid Input


I declare that I have read and fully understood the Privacy Notice and I hereby express my consent to the processing of my personal data by Reply SpA for marketing purposes, in particular to receive promotional and commercial communications or information regarding company events or webinars, using automated contact means (e.g. SMS, MMS, fax, email and web applications) or traditional methods (e.g. phone calls and paper mail).

Fraud detection is a crucial task in various industries such as finance, insurance, and e-commerce. A major problem, however, is the high rate of false positives in alerts. False positives occur when a legitimate transaction is flagged as fraudulent, causing inconvenience to the customer and wasting the resources of the organisation. False positives are caused by various factors such as inaccurate or incomplete data, outdated algorithms, manual intervention, and the lack of contextual information. In addition, fraudsters are getting more sophisticated in their techniques, making it harder for traditional rule-based systems to detect fraud accurately.  

One way to reduce false positives is to implement machine learning algorithms that can learn from the data and improve over time. However, this also introduces new challenges such as the need for large amounts of high-quality data and the risk of model bias. Machine learning models have been implemented more generally in a retrospective context for fraud detection. This means that they are more geared towards reducing alerts associated with false positives rather than creating smarter alerts. Alternatively, another approach is to use network analysis techniques to detect fraud by identifying suspicious patterns of behaviour among connected entities such as customers, merchants, and devices. By analysing the relationships between entities, network analysis can uncover hidden fraud rings and other types of organised fraud.

In network science, a clique is a subset of nodes in a network that are all connected to each other. A clique can be defined as a fully connected subgraph of a larger network, where each node is directly connected to every other node in the clique. The concept of cliques is widely used in social network analysis, where it is used to identify groups of friends or co-workers who are highly connected to each other.

Below are some of the main uses of cliques in fraud detection across various industries:

In the finance industry, cliques can be used to detect fraudulent activities such as money laundering, identity theft, and credit card fraud. By identifying fraud rings, banks and financial institutions can take action to prevent further fraudulent activities and protect their customers.

In the insurance industry, cliques can be used to detect fraudulent claims by identifying networks of individuals who are colluding to make false claims.

In the e-commerce industry, cliques can be used to detect fraudulent activities such as fake reviews, click fraud, etc.

What are Cliques?

How Can It Be Implemented?

Step 1

We first start with the original transaction network. The transaction network is imported into a cloud-based graph database, which is queried to retrieve a temporal slice of the original network.

Step 2

We then perform a network reduction technique and remove potential middlemen and entities which we might not consider suspicious. We then focus our attention on the most densely connected sub-communities within the transaction network - these are the most likely to contain suspicious clique patterns. We are in search of patterns within a transaction network that is suspicious, we can call these patterns heavy cliques and are fully-connected sub-regions of a transaction network.

Step 3

We run the recursive clique detection algorithm to find the suspicious entities. These patterns help identify suspicious entities in the wider transaction networks.

Step 4

Output any suspicious entities into cloud-based storage for further analysis. We can run this process in a batch format every day, week, or month..

The Benefits

1. Graph technology has been shown to perform better in fraud detection than traditional rule-based methods and has been made more available with the rise in computational power available, which allows users to have more precise and smarter alerts..

2. It can also integrate with existing systems, such as rule-based detection options, to provide a comprehensive view of financial transactions, allowing financial institutions to take faster and more informed decisions to prevent money laundering.

3. Rule-based systems can be further refined with clique detection running in parallel thereby reducing false positives and enhancing operational efficiency.

4. Fines amounting to millions of pounds could be mitigated by generating smarter alerts with a solution that costs a fraction of them, thereby reducing overheads.

If you're interested in learning more about utilising network CLIQUE detection for advanced anti-money laundering, please get in touch.

Data Reply is a Reply Group company, a premier AWS partner, offering a broad range of advanced analytics, AI ML and data processing services. We operate across different industries and business functions, enabling our customers to achieve meaningful business outcomes through effective use of data, accelerating innovation and time to value.

Amazon Web Services (AWS) is the world’s most comprehensive and broadly adopted cloud platform, offering over 200 fully featured services from data centres globally. Millions of customers—including the fastest-growing start-ups, largest enterprises, and leading government agencies—are using AWS to lower costs, become more agile, and innovate faster.