Which of the following statements regarding the data analytics approach known as predictive analytics is...

Within CFE investigation guidance on data analysis, predictive approaches are framed as using analytical techniques (including statistical processes and machine learning) to identify patterns and outliers that can indicate fraud risk and, importantly, help detect a fraudulent situation earlier than traditional methods. Predictive analytics is not limited to one specific technology such as layered neural networks (A); neural networks can be one possible technique, but predictive analytics is broader than that. Option B describes robotic process automation (RPA)—software that imitates human actions to automate routine tasks—rather than predictive analytics. Option D describes artificial intelligence in a general sense (systems performing “human tasks” like perception and decision-making), which is also broader and not specific to predictive modeling. The most accurate description is therefore C: predictive analytics uses historical data along with statistical analysis and machine learning methods to build models that reflect patterns/trends and generate projections about likely future outcomes (e.g., predicting which transactions, vendors, employees, or accounts are higher-risk).