Evaluation of feature selection methods

Fighting sophisticated fraud requires using powerful machine models – but how do you build these models?

Raw data is the basic building block of ML algorithms, but on its own it can’t be used to accurately train these models. Instead, it must be refined to “features” – variables or attributes that can be used for analysis, like a name or address. When done well, feature selection helps enhance generalization, reduce training times, increase model interpretability, improve accuracy, and reduce prediction time. However, not all feature selection approaches are created equal.

Read this white paper to learn:

• What the process looks like from taking raw data and refining it into features
• ​Which features are most likely to help us predict fraud
• How to remove any irrelevant or noisy features, leaving us with only the features most relevant for fraud detection