VFlash: Fast, Accurate Visual Planning via Flash Primitive Decoding
u_Unrolled
Easy10 XPoverfitting

Training accuracy 98%, validation 70%: what is wrong?

A gradient-boosted model scores 98% on training data and 70% on validation. What is the most likely diagnosis?

Stuck? Show the solution

Answer: B. Overfitting: regularize, simplify, or get more data

A large gap between training and validation scores means the model memorized the training data. Limit depth, add regularization or early stopping, or get more data.

Learn the idea behind itSupervised Learning — Learning From Labelled Examples