Timeline for How to interpret the bandwidth value in a kernel density estimation?
Current License: CC BY-SA 3.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 29, 2016 at 8:59 | history | edited | Tim |
edited tags
|
|
| Jul 29, 2016 at 8:52 | history | edited | Tim | CC BY-SA 3.0 |
deleted 8 characters in body; edited tags
|
| Jul 29, 2016 at 8:32 | comment | added | Igor | Which bandwidth to choose is also somewhat dependent on the kernel you are using, e.g. for a uniform kernel $$K(x)\propto \mathbf{1}_{[|x|\leq1]}$$ a bandwidth $h=2$ would mean that $K( (x-x_i)\cdot h^{-1})$ would be nonzero around $[x_i-2, x_i+2]$. | |
| Jul 29, 2016 at 8:30 | answer | added | Tim | timeline score: 8 | |
| Jul 29, 2016 at 7:52 | comment | added | Johann | So, the bandwidth value specifies the "range of points" covered on the x axis and the type of kernel specifies its height and shapre. Is that correct? | |
| Jul 29, 2016 at 7:48 | comment | added | Antoine | think of the KDE as a smoothed version of the histogram | |
| Jul 29, 2016 at 7:45 | comment | added | user83346 | You interpret it more or less correctly with that nuance that the correct interpretation depends on the kernel that you choose | |
| Jul 29, 2016 at 7:41 | review | First posts | |||
| Jul 29, 2016 at 7:48 | |||||
| Jul 29, 2016 at 7:38 | history | asked | Johann | CC BY-SA 3.0 |