Zero sample correlation does not necessarily mean there is no linear relationship
Try measuring through the Spearman's Rank Correlation Coefficient - Spearman's Rank Correlation indicator
Thanks, I'll take a look. But you're quoted from here:
The power of the Spearman rank correlation coefficient is slightly inferior to that of the parametric correlation coefficient.
Your indicator shows autocorrelation. And it doesn't seem to count it correctly...
1. Thanks, I'll take a look at it. But you are the one quoted from here:
2. Your indicator shows autocorrelation. And it doesn't seem to count it correctly...
1. Spearman will indeed show a higher correlation value if there is one. Pyroson's strength is that it will only show one if the data are completely identical. Spearman does not require complete identity for a value of 1.
2. Not true.
In fact, it is written in the books that if the RR = 0, it does not mean that the two quantities in question are unrelated.
The link that Rosh gave is exactly Spearman's Rank Correlation Coefficient. That's how it's calculated. If you want to see autocorrelation, it is calculated a little differently, like this https://www.mql5.com/ru/code/8295
Count the correlation with a small period, then count the ratio of the number of high correlation values to the total number of bars. It will be more indicative. By this method, the USDNOK - USDSGD correlation is greater than 0.5 - there is a significant one.
But that's not what we're talking about here. I don't care what the correlation shows if the point is not clear.
My conclusion is that correlation (Pearson's coefficient) is a shitty indicator of the presence of a linear relationship in a sample. Not only does the correlation not show a direct correlation, it also lies.
hrenfx:
I.e. the correlation in this case is the sum of the products of the BP terms divided by the length of the BP.
Why on earth would you do that?
Because the MO is zero and the variance is one.
1. Spearman will indeed show a higher correlation if there is one. Pyroson's strength is that it will only show one if the data are completely identical. Spearman does not require complete identity for a value of 1.
2. Not true.
1. You are confused. Pearson's coefficient shows one for complete identity.
2. True. Autocorrelation is the correlation of BP and its shift. In this case, the Spearman autocorrelation counts.
Yes, it is possible to plot the change in correlation as the sample window moves. And then plot it by MO. This is no longer a correlation, but an average correlation across the window.
Tuna-tuna tapping with a wrench on the helmet. Read my first post in this thread more carefully.
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
Everywhere I read, they write that zero sample correlation means there is no linear (usually forget the word linear too) relationship in that sample. Dickens:
Example of two graphs with zero MO, variance one and zero correlation. That is, the correlation in this case is the sum of the products of the BP terms divided by the length of the BP.
These are the EURUSD and GBPUSD charts over the interval 2010.09.28 13:45 - 2010.09.29 14:15.
If the sample seems small, let's take something larger from the correlation table:
In fact, there is always a linear relationship between any two random variables on a finite sample.
Be careful about interpreting correlations close to zero.