Random Forests

Random forests can also be used to identify likely fraudulent transactions. For example, each transaction in a bank has a series of features such as the deviation from the mean transaction volume of the customer, the time of day, the location, and how these values differ from that customer's usual habits. This allows a bank to build a sophisticated model to predict the likelihood of a given transaction being fraudulent. If the probability of fraud exceeds a threshold, such as 50%, the bank can take action, such as freezing the card.