#KFold
LightNews LightNews
tjmahr.com
brms::kfold(model, group = group_var)
December 4, 2024 at 8:01 PM
successchukwuka.bsky.social
Eight weeks. 22 classes. Two left. The finish line is right there.

Follow the journey 👇

#DataScience #MachineLearning #LinearRegression #KFold #LearningInPublic #Tech4Africans #TechCrush #CareerTransition
May 5, 2026 at 9:28 PM
ctrlalttim.com
I've organized all the code I wrote to do workshops and classes about species distribution models as a #JuliaLang package, and it should be online in a few days. Look at how nice the code is to use!!!
Two maps of Switzerland, one showing the prediction of a species distribution model, and the second one showing the uncertainty # Create the model sdm = SDM(MultivariateTransform{PCA}, NaiveBayes, layers, presencelayer, bgpoints) # Cross-validation + variable selection folds = kfold(sdm); forwardselection!(sdm, folds; verbose=true) # Prediction of probability prd = predict(sdm, layers; threshold = false) # Uncertainty via bootstrap ensemble = Bagging(sdm, 20) train!(ensembe) unc = predict(ensemble, layers; consensus = iqr, threshold = false)
September 23, 2024 at 3:42 PM
gemmaeling.bsky.social
But stress genes are only part of the picture. What about genes involved in development, anatomy or biochemistry? To look into this, we trained a machine learning model to distinguish between aquatic and terrestrial species based on the number of genes in each OG.

8/14
Simplified workflow of machine learning analysis. Input OGs are fed into XGBoost classifier with Kfold stratified cross validation and dRFE. Output is separation between aquatic and terrestrial, and top OGs
September 3, 2026 at 10:40 AM
pytorch.kr
그리드 서치
(by Jonna_difficult님)

https://d.ptln.kr/3883

#ai #파이토치
그리드 서치
파이토치로 딥러닝 모델 구축하는 과정에서 하이퍼파라미터 튜닝에 대해서 질문 있습니다. Optuna, GridSearch 이러한 하이퍼파라미터 튜닝을 통해서 최적화 파라미터를 구하게 되었다면 이 값을 갖고 다시 학습 시키는게 맞는 방법인가요? GridSearch CV 를 사용하면 kfold값을 5로 지정 하면서 파라미터 찾는 과정에서 이미 학습이 진행 된거여서 여기 나온 모델을 바로 사용하는게 맞는건지 모르겠습니다
d.ptln.kr
March 29, 2024 at 5:38 AM
dailyqiitatrends.bsky.social
今日のQiitaトレンド

KFold最適化テクニック～StratifiedKFoldを超えろ
この記事では、機械学習におけるKFoldのseedを探索することでCVスコアを向上させる手法を紹介している。
Stratified KFoldではFold間のデータ分布の偏りを抑えるが、さらにseedを調整することでCVスコアに最適な「秘孔」を見つけられる可能性があるという。
この手法はデータ量が少ないコンペで有効だが、実務への応用は限定的であることを示唆している。
KFold最適化テクニック～StratifiedKFoldを超えろ #機械学習 - Qiita
TL;DRKFoldのseedをたくさん試して手元のCVスコアを上げるのは有効なテクニックであるそれが有効であることの数学的証明はないが、Stratified KFoldという機能の存在が傍証に…
qiita.com
September 4, 2024 at 10:18 PM
fusaroli.eurosky.social
existing kfold/kfold_predict() functions)
November 13, 2024 at 8:55 PM
fusaroli.eurosky.social
is there a way to quickly run predictions from kfold crossvalidated #brms models on an external dataset? https://discourse.mc-stan.org/t/kfold-predictions-on-an-external-test-set/28989?u=fusaroli (I can write a stratification + loop script, but something off-the-shelf using the...
Kfold predictions on an external test set
For a course I’m teaching I want to showcase the changes ...
discourse.mc-stan.org
November 13, 2024 at 8:38 PM
bluemendal.bsky.social
descobrindo HOJE a função group kfold do scikit learn que faz basicamente algo que eu fazia na unha a pelo menos 7 meses

☠️
October 3, 2024 at 10:40 PM
axjack.bsky.social
kfoldで出てきた一時的なモデルは評価のために利用するけど、最終的には全訓練データを用いたモデルを使う、のが一般的なのかなぁ。
May 20, 2024 at 11:59 PM