Linear and Ranked Linear Regression

Left Image Shows original EUR/USD values and the corresponding Regression line based on the original values. Right Image Shows the ranked values and ist corresponding ranked Regression line. Values are OK, double checked with R and so forth even though the look pretty similar. This is due to the abscence of "real" outliers…

Since timeseries data such as EUR/USD is not normal distributed, I used Ranked Linear Regression to fit the line and to calculate a p-value using bootstrapping. All works fine apart from plotting the ranked Regression line in the Barchart itself as the fitted values for the ranked Regression line lie between 3.0 and 8.0 approximately. As the original data lies between 1.13860 and 1.13920 you are not able to plot the ranked Regression line.

Any ideas how to fix this? I thought About some scaling but not yet really come up with a good solution. Not sure, whether this is possible at all as from original to ranked data you lose Information...


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.