Credit card fraud is a major concern in today's world, as it involves the illegal use of credit cards to obtain goods or services. The process of credit card fraud often involves "laundering" dirty money, making it difficult to trace the source of funds. With the large volume of financial transactions happening globally, It might be difficult to identify credit card theft. In the past, anti-fraud suites were introduced to detect suspicious activity on individual transactions, but they were not effective in detecting fraud across multiple bank accounts. To overcome this challenge, we propose using machine learning techniques, specifically the 'Structural Similarity' method, to identify common attributes and behavior across multiple bank account transactions. It can be challenging to identify credit card fraud from huge datasets, so we also suggest utilizing case reduction techniques to shrink the input dataset and then looking for pairs of transactions with similar characteristics and behaviours.