Predictive analytics

The need to predict outcomes in medicine is growing as providers, payers, and patients strive to maximize outcomes, minimize costs and incorporate shared decision making into practice. Coupled with ever more available databases and improved computational ability, one can expect predictive models to flourish [1,2]. Machine learning algorithms offer the promise of improved modeling particularly with nonlinear data and large, complex datasets. However, with these more flexible approaches comes the risk of overfitting one's data thereby making them less useful for generalized usage [3].

Source: The Spine Journal - June 30, 2020 Category: Orthopaedics Authors: Joseph H. Schwab Tags: Commentary Source Type: research