Skip to main content
. 2021 Feb 16;11:3932. doi: 10.1038/s41598-021-83141-z

Table 1.

Differences in image features and model complexity when comparing standard GP and transductive learning GP models.

Transductive Learning GP model Standard GP model
Selected image texture features

1. T2.Information.Measure.of.Correlation.2_Avg_1

2. T2.Angular.Second.Moment_Avg_1

3. T2.Kurtosis

4. rCBV.Contrast_Avg_1

1. T2.Difference.Entropy_Avg_3

2. T2.Contrast_Avg_1

3. T2.Entropy_Avg_1

4. SPGRC.Sum.Variance_Avg_3

5. SPGRC.Gabor_Std_0.4_0.1

6. rCBV.Angular.Second.Moment_Avg_1

7. rCBV.Difference.Variance_Avg_1

8. rCBV.Kurtosis

9. EPI.Gabor_Mean_0.4_0.1

10. EPI.Angular.Second.Moment_Avg_1

11. FA.Angular.Second.Moment_Avg_1

12. FA.Skewness

13. FA.Difference.Variance_Avg_3

14. FA.Entropy_Avg_1

15. FA.Sum.Variance_Avg_1

16. FA.Information.Measure.of.Correlation.1_Avg_1

17. FA.Range"

The transductive learning model, which prioritized model uncertainty during the training process, comprised fewer image features and lower complexity compared with the standard GP model.