Correlation is used in Cause and Effect Analysis . When we start the analysis without the knowledge of cause-variables and effect-variables, we find not only the relation between causes and effects but also the relation between effects and effects.
Below is the analysis of the relation between effects and effects. This is the method to use suspected correlation effectively.
When X is the profit of A and Y is the profit of B, if there is positive correlation, it tells us the win-win relationship. If there is negative correlation, it tells us the trade-off relationship.
This method is used to compare two measurement machines. We compare Y and Y for each X. In this method, values of X is not used to calculate. So if X is quality data, this method can be used.
Concept of this method is used for Standard SN Ratio of Quality Engineering .
In the prediction analysis, if the prediction function is good, the plot of predicted value and measured value is on the line. So the ability of prediction (fitting) is evaluated by correlation coefficient.
We compare measured Y and predicted Y'. Y' is calculated by the prediction function (for example, made by regression analysis ).