Përdorimi i Decision Trees dhe të dhënat për të marrë vendime
summary
Shows a concrete pipeline: feature engineering from web and server logs, training CART decision trees with scikit-learn, pruning by information gain and early stopping. Reports results on a ~10k-record dataset with 70–80% accuracy and lower false positives than a baseline logistic regression. Ships a small Angular demo wired to a REST endpoint that predicts next user action and demonstrates reduced misclassification in live input. Explains where trees fail and which features actually drive splits.
AI-written summary · about
related talks
40mCART: Not only Classification and Regression Trees
21mBuild, Measure, and Machine Learn with Angular
35mUnderstanding, Simplicity, And Decision Trees
02h 03mFull stack AI design and development
45m"Breaking black-box machine learning"
44mUnderstanding Random Forests
27mData-driven challenges in AI: scale, information selection, and safety
42mJust-So Stories for AI: Explaining Black-Box Predictions