Individual Player Value (IPV) is a Plus/Minus model, created by Talking Practice blog. IPV employs a robust machine learning model by running the standard RAPM procedure and regressing that to the Bayesian prior.
Individual Player Value incorporates boxscore, play-by-play, video tracking stats, and a variety of information such as player demographics.
With standard regression techniques, you worry too much about overfitting the dataset. However, a machine learning model performs better out of sample than more plain regression-based models (especially with regard to defense).