Project 3 · Classical ML0/5 completeClassifier Lab
A score, a threshold, a loss, and a gradient — the learner fits a logistic classifier from scratch on the Project 1 split and Project 2 features.
requires artifacts from: metrics-lab
- L1Score Functiondefine the contract · ~12 minstart →
- L2Thresholded Predictionpreserve an invariant · ~14 minlocked
- L3Sigmoid + Log Lossmeasure behavior · ~16 minlocked
- L4Gradient Descentbound a failure · ~22 minlocked
- L5Trainable Artifactmake an operational decision · ~28 minlocked
Artifacts this project producesL1 · Score Functionoutputs: classifier-lab:l1-score-function:artifact
L2 · Thresholded Predictionoutputs: classifier-lab:l2-thresholded-prediction:artifact
L3 · Sigmoid + Log Lossoutputs: classifier-lab:l3-sigmoid-logloss:artifact
L4 · Gradient Descentoutputs: classifier-lab:l4-gradient-descent:artifact
L5 · Trainable Artifactoutputs: classifier-lab:l5-trainable-artifact:artifact