Logistic Regression for Classification
Model yes or no outcomes, read odds ratios and choose thresholds with intent
A taste of a lesson
My coefficient for 'opened a support ticket' is 0.7. Does that mean a 70 percent higher chance of churn?
Not quite. The coefficient lives on the log odds scale. exp(0.7) is about 2.0, so opening a ticket roughly doubles the odds of churn, holding the other features constant. Doubling odds is not doubling probability. If a customer's churn probability was 10 percent, odds of 1 to 9, doubling gives 2 to 9, which is about 18 percent. Quick exercise: what probability do you get if the starting probability is 50 percent and the odds double?
Written by the teacher as an example. In your lesson the tutor answers your own questions, and like any AI it can be wrong.
What you will be able to do
- Explain how the sigmoid turns a linear score into a probability
- Translate coefficients into odds ratios and plain sentences
- Describe how log loss trains the model and why regularisation matters
- Choose a decision threshold from the costs of different errors
- Recognise perfect separation and multiclass setups
Lesson plan
- 1 Why a straight line fails See why linear regression gives impossible probabilities for binary targets. Start
- 2 Odds and log odds Move fluently between probability, odds and log odds. Start
- 3 Training with log loss Understand what the model optimises and why confident mistakes hurt. Start
- 4 Reading coefficients Interpret coefficients as odds ratios with care. Start
- 5 From probability to decision Pick a threshold based on the costs of false alarms and misses. Start
- 6 More than two classes Extend logistic regression to multiclass problems. Start
Try asking
About this tutor
An intermediate tutor for learners who know linear regression and now need to predict categories. You will see why a straight line fails for probabilities, how the sigmoid turns a linear score into a probability, and how log loss trains the model. Lessons cover reading coefficients as odds ratios, scaling and regularisation, separating classes perfectly, multiclass extensions and choosing a decision threshold from the costs of errors. You practise on small cases such as loan defaults and support tickets, explaining outputs the way you would to a colleague.
Reviews
Students can review a tutor after a paid lesson. Nobody has yet.
About the teacher
Classical machine learning models, worked through on paper before any code
9 tutors 379 lessons taught Sample
I teach the classical machine learning models: regression, trees, ensembles, nearest neighbours, support vector machines and naive Bayes. My background is in applied analytics, where I spent a lot of time explaining to colleagues why a model made a particular prediction, and that shaped how I teach. I like to start with a tiny dataset you can hold in your...
See Kavya's profile and tutorsMore like this
Other tutors on the same or nearby topics.