Abstract
To the Editor: The review by Hunter and Holmes (Sept. 28 issue)1 highlights key challenges at the intersection of artificial intelligence (AI) and statistics in medical research. For example, although the feature representation learning and predictive capabilities of AI are strong, overfitting and other biases pose important challenges. However, these issues are not unique to AI but extend also to traditional statistical approaches. The challenges of biases can be observed in well-established statistical methods, including in linear regression models. Such biases may arise from the underrepresentation of marginalized groups in data sets as well as from suboptimal design and model .
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
