// the find
atilsamancioglu/MachineLearningNotebooks
A flat collection of 30+ Jupyter notebooks walking through classical ML algorithms one by one — linear regression through XGBoost, PCA, clustering, and anomaly detection. Each notebook is paired with a matching CSV dataset. Aimed at ML beginners working through sklearn-based supervised and unsupervised methods.
Clean one-topic-per-notebook structure makes it easy to jump straight to whichever algorithm you need. Dataset variety is decent — fraud detection, cyber attack classification, and seismic activity give slightly more interesting contexts than the usual iris/titanic loop. Coverage of the boosting family (AdaBoost, GradientBoosting, XGBoost, LightGBM) is more complete than most intro repos. Notebook 30 covering model serialization with pickle is a practical detail many tutorial sets skip.
No README and no explanatory prose visible from the tree — if the notebooks themselves are mostly code cells with minimal markdown, you're reading code without knowing why decisions were made. The flat file structure with 60+ files in one directory gets unwieldy fast. Nothing here touches neural networks, transformers, or any modern deep learning — the content stops at 2019-era sklearn. Last push was March 2026 but the content itself hasn't evolved beyond what was available in scikit-learn 1.x.