/coursera-machine-learning

My solutions to Stanford's Machine Learning course on Coursera

Primary LanguageJupyter NotebookMIT LicenseMIT

Coursera Machine Learning

These are my solutions to the exercises for Stanford's Machine Learning course on Coursera: https://www.coursera.org/learn/machine-learning/home/info

On some of the exercises, I use Matlab (or Octave) as suggested by the course's teacher Andrew Ng. For some of them, I provide a solution in Python + numPy.

I think that using Jupyter Notebooks for these exercises is a great idea, they work nicely with Octave and Python. For details on how to run Jupyter Notebooks with Octave, see: https://github.com/Calysto/octave_kernel

Exercises:

  1. Linear Regression: Predict profit for a food truck. Octave, Python
  2. Logistic Regression: Octave
    1. Admission of students
    2. QA for micro chips
  3. Multi class classification: Recognition of handwritten digits. Octave
    1. One-vs-all logistic regression
    2. Applying a pre-learned neural network
  4. Training a neural network: Recognition of handwritten digits. Octave
  5. Bias and Variance Problems. Octave
    1. Linear regression
    2. Polynomial features
    3. Regularization
    4. Learning curves
    5. Using cross validation data and test data