KDE를 구하기 위한 python package
1. scipy
2. scikit-learn
3. statsmodels
이렇게 세 가지가 있는지 몰랐다....
그래도 이제는 알았지.
무려 7년전 포스팅이긴 하지만 내 궁금증이 좀 해결되었던 참고블로그:
https://jakevdp.github.io/blog/2013/12/01/kernel-density-estimation/
Kernel Density Estimation in Python | Pythonic Perambulations
Notice the two regions of interest: for very small bandwidths, kernels with a hard cutoff (tophat, linear) out-perform kernels with a broad taper (gaussian, exponential). And tapered kernels which fall off more quickly (gaussian, with $p \sim \exp(-d^2)$)
jakevdp.github.io
+ (20210709) 비교해 놓은것들 종종 있네
https://gist.github.com/daleroberts/7a13afed55f3e2388865b0ec94cd80d2
Two-dimensional kernel density estimate: comparing scikit-learn and scipy
Two-dimensional kernel density estimate: comparing scikit-learn and scipy - kde.png
gist.github.com