Skip to main content
9 events
when toggle format what by license comment
Jul 6 at 11:44 audit First answers
Jul 6 at 11:56
Jul 6 at 6:55 comment added YiFan Tey @XanderHenderson I don't think that explanation is correct. The distinction is between probabilities and probability densities; the point is that $f(x)$ is not a probability (and hence cannot be written in percentages) --- which would indeed be dimensionless --- but is instead a probability density.
Jul 6 at 6:33 comment added DannyNiu @XanderHenderson Not to insist on arguing, but for most of the time, I just assume the percentage symbol % has the unitless value of 0.01. It's multiplied with a human-friendly 2-digit number (occasionally 1-digit or 3-digit) to indicate progress or fraction. This way, "%" is generic, and the "amount per 100" can be per 100 whatever the author wishes.
Jul 5 at 19:36 comment added Prem Like I said in my answer , that is not the only way to resolve this issue , @XanderHenderson , there are authors who will use things like "Let $x$ be the length" where $x$ must itself be length units , while there are other authors who will use things like "Let $x$ be the length in meters" where the real number $x$ itself is unitless. With that way , it is then more meaningful to consider $E[x^2+x^3]$ with "Unitless Percentages" , otherwise , we can not add area to volume.
Jul 5 at 18:54 comment added Xander Henderson @DannyNiu Remember that "percent" is amount per 100. Per 100 whats? Oh, per 100 of whatever $x$ is. Hence if you are thinking of $f(x)$ as a percentage, then it has units of $1 / (100[x])$.
Jul 5 at 10:33 history edited Abezhiko CC BY-SA 4.0
deleted 5 characters in body
Jul 5 at 7:38 comment added DannyNiu Thanks. The unit of the density function is what I've missed. Too often $f(x)$ was thought in percentage - i.e. unitless.
Jul 5 at 7:36 vote accept DannyNiu
Jul 5 at 7:34 history answered Abezhiko CC BY-SA 4.0