Significance tests or confidence intervals: Which are preferable for the comparison of classifiers?

Mathematics

• Significance Test 100%
• Classifier 85%
• p-Value 82%
• Effect Size 73%
• Confidence interval 63%
• Test Set 25%
• Hypothesis Test 21%
• Machine Learning 20%
• Small Sample 18%
• Reasoning 18%
• Null hypothesis 18%
• Uncertainty 16%
• Evaluation 13%
• Performance 11%

Engineering & Materials Science

• Classifiers 53%
• Machine learning 18%
• Uncertainty 15%