Table 3 DFNN model performance in Q1 without k-fold.

From: Path-based evaluation of deep learning models for solving inverse kinematics in a revolute-prismatic robot