Selecting predictive features to input into a model is key to ensuring that the input data is not noisy and is time-effective in cases where the original number of features or dataset are large. In this talk I will present a survey of key feature importance packages and explain their strengths and weaknesses, and I will present an in-house open-source feature importance package called LOFO (leave-one-feature-out) and its fast approximation (FLOFO, or Fast LOFO). The LOFO importance calculates the importances of a set of features based on a metric of choice, for a model of choice, by iteratively removing each feature from the set, and evaluating the performance of the model, cross-validated, based on the chosen metric.

Bio: Rafah El-Khatib is a data scientist at ING Bank, working within the Financial Markets and Advanced Analytics departments on machine learning applications in financial investments and trading. She received her B. Eng in Electrical and Computer Engineering from the AUB, Lebanon, where she did research on software verification and signal processing, and her Ph.D. in Computer and Communication Sciences from EPFL, Switzerland. Her research interests include the design and analysis of coding systems with a focus on graphical models, as well as signal processing more generally, and machine learning.