What are best practices around transforming correlation measurements? For my application, I need a figure greater than zero while my input ranges from -1 to +1 inclusive. The application is financial portfolio correlations, if it helps.
Transforming correlations
0
$\begingroup$
statistics
linear-transformations
finance
correlation
-
0I thought about adding 2 to the correlation measurement but it seems to be a rather crude approach – 2017-01-12
-
0It's going to be a denominator in a ratio. The numerator is positive. The output must be positive, and a zero makes the ratio undefined. So I need to transform the correlation measurement to greater than zero. – 2017-01-12