Kurtosis moment is the fourth moment of profile amplitude probability function and corresponds to a measure of surface sharpness. Even than negative value ?
Table is attached with this question.
It means that you have a platykurtic distribution (the tails are heavy, the distribution has broad shoulders, the peak is broader and wider than if kurtosis was a larger value.). At least that is what it indicates in a symmetrical distribution. However, specifying a mean, standard deviation, skewness, and kurtosis is not sufficient to uniquely define a distribution. So what it is really telling you is that it is less likely that your data are normally distributed (though this value alone is insufficient to reject the null hypothesis that the data are normally distributed).
I assume that you are using the Pearson's moment kurtosis.
You might check the user manual for whatever piece of equipment gave these results. It is possible that these values are used as a diagnostic to see if the equipment is performing properly, or it could be a quality control issue (wild guess).
Reading this may help: http://en.wikipedia.org/wiki/Kurtosis.
It means that you have a platykurtic distribution (the tails are heavy, the distribution has broad shoulders, the peak is broader and wider than if kurtosis was a larger value.). At least that is what it indicates in a symmetrical distribution. However, specifying a mean, standard deviation, skewness, and kurtosis is not sufficient to uniquely define a distribution. So what it is really telling you is that it is less likely that your data are normally distributed (though this value alone is insufficient to reject the null hypothesis that the data are normally distributed).
I assume that you are using the Pearson's moment kurtosis.
You might check the user manual for whatever piece of equipment gave these results. It is possible that these values are used as a diagnostic to see if the equipment is performing properly, or it could be a quality control issue (wild guess).
Reading this may help: http://en.wikipedia.org/wiki/Kurtosis.
it seems that the kurtosis values you get correspond to k-3, so positive values correspond to leptokurtotic behavior (heavy tails, sparse signal), zero to Gaussian density, and negative to platokurtotic behavior, meaning that relative large values in amplitude may be as likely or more than close to zero values (a dense or "anti-sparse" response).
A normal distribution has zero kurtosis, a Student distribution with at least five degrees of freedom has positive kurtosis, and a uniform distribution has a negative kurtosis.
The normal distribution has zero kurtosis only if you include the -3 in the formula.
Negative (excess) kurtosis means that the outlier character of your data is less extreme that expected had the data come from a normal distribution.
Positive (excess) kurtosis means that the outlier character of your data is more extreme that expected had the data come from a normal distribution.
Zero (excess) kurtosis means that the outlier character of your data is similar to what is expected had the data come from a normal distribution.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics