Zero sample correlation does not necessarily mean there is no linear relationship - page 41
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Heh, not obvious. For Pearson's QC it doesn't matter whether the rows are positive or negative...
O.K. Let's assume that it does not matter. Then the appearance of the I(0) correlation matrix distribution will be approximately the same as with I(1). Check. Let us take 100 I(0). Let's construct a correlation matrix of these I's with each other. Then construct a histogram of the frequencies of the most frequent values:
We see a classical normal distribution centered around zero - all right, because 100 rows are completely independent of each other. Rarely does the correlation between rows reach +/- 10 percent.
Now take 100 series and integrate them. The output will be a classical random walk of the form I(1). We construct a correlation matrix for these series and then the same distribution histogram:
The distribution is collapsed. Values of -0.5 and +0.5 repeat as often as values of 0.0. The CC becomes a meaningless indicator, as any other number can fall out with the same probability, though there is no dependence between the rows reliably.
Now take 100 BP of type I(1), but add value 100 to each of them. Because of the small variance, this is a significant number for these series. Thus all 100 BPs will be in the positive zone > 0. We look at the histogram:
Indeed, nothing has changed compared to the previous graph. But this does not change the essence and the hypothesis remains valid: I(1) series cannot be used to calculate QC.
That's not the way to do it! You have to calculate the logarithm of the price, then the first differences, then take the logarithm from them, and then calculate the correlation.
Ha-ha-ha!
...
Not zero, but 'no value'. That's why you can get a correlation of kotir with Saturn's rings, and by the same token, nose problems.
...
That's not the way to do it! You have to calculate the logarithm of the price, then the first differences, then take the logarithm from them, and then calculate the correlation.
Ha-ha-ha!
That's not the way to do it! You have to calculate the logarithm of the price, then the first differences, then take the logarithm from them, and then calculate the correlation.
Ha-ha-ha!
Limitations of correlation analysis:
The total values of all factor and outcome variables must follow a multivariate normal distribution.
Wiki
I think my understanding is that QC only works for NR of SV? In real-world series, even the first difference is not a NR.
Limitations of correlation analysis:
The total values of all factor and outcome variables must follow a multivariate normal distribution.
Wiki
I think my understanding is that QC only works for NR of SV? In real-world series, even the first difference is not a NR.
...
wiki
...
Ibid:
Often the luring simplicity of correlation studies encourages the researcher to draw false intuitive conclusions about the existence of a causal relationship between pairs of attributes, whereas correlation coefficients only establish statistical relationships. Looking at fires in a given city, for example, one might find a very high correlation between fire damage and the number of firefighters involved in suppressing a fire, and the correlation would be positive. It does not follow, however, that "more firefighters leads to more damage," and it makes even less sense to try to minimize fire damage by eliminating fire brigades.[5] At the same time, the lack of correlation between the two quantities does not mean that there is no relationship between them.
O.k. Suppose it does not matter. Then the kind of distribution of the correlation matrix ...
The type of the correlation matrix distribution depends on the properties of both series and the relationship between them, i.e. it does not have to be the same for all possible series... For SB it is one, for some solar flares another...