Effector is a Python-based machine learning library that addresses the limitations of global feature effects methods by providing regional feature effect methods. It aims to reduce aggregation bias and increase interpretability and trustworthiness of machine learning models. Effector offers a range of global and regional effect methods, making it easy for users to compare and choose the most suitable method for their application. It is accessible, easy to use, and encourages collaboration and innovation.
54 Impressions