Krpex
Create account
Machine learningintermediate

Machine Learning Fundamentals

Understand and build machine learning models with Python and scikit-learn, from linear regression to model evaluation.

Lessons
13 lessons in 4 modules
Study time
about 7 hours
Duration
~6 weeks at your own pace

Free plan: counts as one of your 2 new courses this month. Plans

What you'll learn

  • Explain supervised and unsupervised learning with examples
  • Train regression, classification and clustering models in scikit-learn
  • Evaluate models properly and avoid overfitting and data leakage
  • Complete an end-to-end ML project

Syllabus

  1. Module 1The big pictureWhat ML is and how projects work.3 lessons
    1. What machine learning isDistinguish ML from rule-based programming with examples.25 min
    2. The ML workflowList the steps from data collection to deployment.30 min
    3. Working with data in pandasLoad, explore and clean a dataset with pandas.30 min
  2. Module 2Supervised learningLearning from labelled examples.4 lessons
    1. Linear regressionTrain a model to predict a number and interpret its coefficients.30 min
    2. Logistic regression and classificationPredict categories and read probabilities.30 min
    3. Decision trees and random forestsTrain tree-based models and explain their decisions.30 min
    4. Feature engineeringEncode categories, scale features and create useful new ones.30 min
  3. Module 3Evaluating modelsMeasure honestly.3 lessons
    1. Train/test split and cross-validationEstimate how a model will perform on new data.30 min
    2. Metrics: accuracy, precision, recall and RMSEChoose the right metric for the problem.30 min
    3. Overfitting, regularisation and tuningDetect overfitting and tune hyperparameters.30 min
  4. Module 4Beyond the basicsUnsupervised learning and responsible ML.3 lessons
    1. Clustering with k-meansGroup similar data points without labels.30 min
    2. Bias, fairness and data leakageSpot common ways ML projects go wrong.30 min
    3. Project: predict house prices end to endBuild, evaluate and explain a complete model.50 min