start

Search

You can find the results of your search below.

Statistics / Machine Learning
8 Hits, Last modified:
$\sum $ (% of non-events - % of events) * WOE * Random forest can detect non-linear relationship very we... variables via Information Value and using them in random forest model might not produce the most accurate ... ure 와 붙인 뒤에 feature importance를 알 수 있는 의사결정나무 기반의 Random forest, XGBoost 등의 모델로 훈련시켜 shadow features 의 fea... (concat) 4) Feature importance를 잴 수 있는 의사결정나무 기반의 Random forest나 XGBoost 등의 모델을 활용하여 학습합니다. 5) 학습 결과 나온 fe
Tree-based Model, Bagging
7 Hits, Last modified:
=== Tree-based Model, Bagging ====== [[blog:easy_random_forest]] ===== Decision Tree ===== ====Regressio... \rho)}\over{B}} \sigma^2 $ 하지만 각각 독립인 것은 어떻게? -> Random Forest ===== Random Forest ===== 트리들의 correlation을 줄여줌. 상관성을 제거(decorrelate) 간단한 방법으로 배깅된 트리들보다 더 나은 성능을... 큰 경우 OOB오차는 LOOCV 오차와 사실상 동일 ====중요도==== * by random permutation (p번 더 계산) : OOB prediction (j-th 변수 r
Convergence of RV and Testing
4 Hits, Last modified:
== [[blog:easy_clt_survey]] ==== Convergence of Random Variable ==== ===The Weak Law of Large Numbers (... infty} {1\over{\sqrt{2\pi}}} exp(-x^2/2) dx $ => random variable 그 자체가 아닌, probablity statements 가 근사적으로 ... ly observed. H0이 True일 때, p-value 는 Unif(0,1) 에서 random draw와 같음. H1이 True일 때, p-value의 distribution은 0에... heta$, since $\theta $ is a fixed quantity, not a random variable. * 1. if i repeat the experiment ove
A/B test
1 Hits, Last modified:
control/treatment, 지표, bayesian/frequentist * Random unit : user/session/page * Target unit : 전체 /
Probability Distributions
1 Hits, Last modified:
{k+r-1 \choose r} p^k (1-p)^r$$ where $X$ is the random variable representing the number of trials until
Review on BPR: Bayesian Personalized Ranking from Implicit Feedback, Rendle et al.
1 Hits, Last modified:
- correspond to the MLE for normally distributed random variables. - But item prediction is actually no
  • Last modified: 2025/07/07 14:12