Cortex AI logo placeholderCortexLabAI · learning · engineering
Model Atlas · Linear models

Logistic Regression

A probabilistic classification baseline that is transparent, fast and surprisingly competitive.

core mathematical viewp(y=1|x)=σ(wᵀx+b)
Mental model

Understand it before memorizing it.

Learn a weighted score then pass it through a sigmoid to obtain a probability.

Best fit

Where this model earns its place

Binary classification

Start with a simpler baseline then compare this model using the same evaluation split and operational constraints.

Interpretable baselines

Start with a simpler baseline then compare this model using the same evaluation split and operational constraints.

Calibrated probability starting point

Start with a simpler baseline then compare this model using the same evaluation split and operational constraints.

Strengths and limits

Trade-offs matter more than popularity.

Strengths

✓ Fast

✓ Explainable coefficients

✓ Small memory footprint

Limitations

△ Linear boundary

△ Interaction engineering may be needed

Evaluation

Metrics to watch

PR-AUCInterpret with the product objective and error cost.
ROC-AUCInterpret with the product objective and error cost.
Log lossInterpret with the product objective and error cost.
Calibration errorInterpret with the product objective and error cost.
Production checklist

Before it reaches real users

  1. 01

    Standardize only when useful

    Document the assumption and instrument the condition so regressions can be detected.

  2. 02

    Monitor calibration

    Document the assumption and instrument the condition so regressions can be detected.

  3. 03

    Choose thresholds from business cost

    Document the assumption and instrument the condition so regressions can be detected.