InData Science CollectivebyIlia Ekhlakov·3d agoWhy DART wins your Optuna slice plot and loses the benchmarkThree things an Optuna slice plot hid about DART, checked on 131 small datasets with LightGBM and XGBoosA response icon1A response icon1
Burcu İlayda Şentürk·Oct 20İyzico Üye İş Yeri Bazında 92 Günlük Ödeme Hacmi Tahmini: COVID-19 Anomalisiyle Başa ÇıkmakZaman serisi tahmininde en zor problemlerden birini pandemi dönemindeki yapısal değişimleri farklı stratejilerle nasıl ele aldım?A response icon1A response icon1
DESHAPRIYA D.M.K.N·Sep 28A researcher’s journey through 6 notebooks, 11 models, and one insight that mattered more than any…The Problem Nobody Talks About in Time-Series Forecasting
Taha Soylu·Sep 24Tree Based ML ModelsIn this section, I am going to mention Machine Learning models using Tree structure.
Sathwik Pinninti·Sep 14Decision Trees with LightGBM style Quantile binning and Histogram optimisation: How I built…I think there were times when I encountered few challenges in terms of increasing complexity of whatever I start to explore, making me…
Zobir ZEGHOUD·Aug 13Why LightGBM Kept Crashing the Moment I Imported MLflow (on Windows)An access violation with no useful stack trace, three wrong hypotheses, and the transitive dependency actually at fault
Zobir ZEGHOUD·Aug 13I Had 100 Rows Per Country. Here’s Why I Didn’t Reach for a Foundation ModelChronos and TimeGPT are one pip install away. The data decided against them before the hype did.
Alejandro Gómez Ruiz·Jul 28Stop training a thousand LightGBMs. It’s the worst thing you can do.Part three of a series putting common ML habits to the test. LightGBM’s one selling point is speed, so training a thousand of them to hunt…
Pratik Bothra·Jul 24My First Kaggle Notebook: What Building a “Simple” House Price Model Actually Taught MeA few weeks ago, I did something a lot of people learning machine learning eventually do: I opened Kaggle, picked the House Prices…
