Hello, at the beginning of the lecture we see two power laws that fit the in-degree distr. and out-degree distr.
It is said that skewed = small lambda parameter, could you explain why please ?
I saw on internet what "skew" means for a normal distribution (one tail more heavy than the other) but I cannot understand this definition in the power law context.
The 'skew' concept for the power law is similar to that for the normal distribution. If the tail is heavier, then we can say it is more skewed. In this two plots, the y-axis is the same but the degree range on x-axis in in-degree plot is wider. So we say that 'in-degree is more skewed'. At the same time, we also notice that the parameters of the in-degree distribution are smaller after fitting.
Skewed power law
Hello, at the beginning of the lecture we see two power laws that fit the in-degree distr. and out-degree distr.
It is said that skewed = small lambda parameter, could you explain why please ?
I saw on internet what "skew" means for a normal distribution (one tail more heavy than the other) but I cannot understand this definition in the power law context.
Thank you for your time,
Sophie
Hi Sophie,
The 'skew' concept for the power law is similar to that for the normal distribution. If the tail is heavier, then we can say it is more skewed. In this two plots, the y-axis is the same but the degree range on x-axis in in-degree plot is wider. So we say that 'in-degree is more skewed'. At the same time, we also notice that the parameters of the in-degree distribution are smaller after fitting.
Best,
Junze
Add comment