Definition
Pearson Correlation
Measures linear relationship between two variables, -1 to +1. Dangerous when used to validate derived scores against themselves.
definitionstatisticsmetrics
Measures linear relationship between two variables, -1 to +1. Dangerous when used to validate derived scores against themselves.
Pearson correlation measures the linear relationship between two variables on a scale from -1 (perfect inverse) to +1 (perfect positive). A score of 0.95 between your model predictions and ground truth sounds impressive, but if both variables are derived from the same input, you are measuring how well your formula agrees with itself. This tautological trap is why DQI was rebuilt from Pearson-on-derived-scores to objective verifiable properties.