Skip to main content
. 2014 Oct 2;85(4):597–603. doi: 10.2319/052514-370.1

Figure 3.

Figure 3. Training and validation errors in absolute values (mm). (Top) Training errors did not show a statistically significant difference according to the validation methods. (Bottom) In general, validation results produced by three cross-validation methods did not demonstrate a significant difference among them. However, the simple validation showed significantly larger absolute errors than the cross-validation methods did, especially for several soft tissue responses in the vertical direction.

Training and validation errors in absolute values (mm). (Top) Training errors did not show a statistically significant difference according to the validation methods. (Bottom) In general, validation results produced by three cross-validation methods did not demonstrate a significant difference among them. However, the simple validation showed significantly larger absolute errors than the cross-validation methods did, especially for several soft tissue responses in the vertical direction.