Exercise 1: Introduction to Machine Learning

Setup

Make sure that you are inside of the intro_ml folder by using the cd command to change directories if needed.

Run the setup script to create the environment for this exercise and download the dataset.

source setup.sh

Launch a jupyter environment

jupyter lab

...and continue with the instructions in the notebook exercise.ipynb.

Exercise

Part A: The Linear Classifier

We will implement a basic linear classifier from scratch and train it to predict the cell cycle on a flow cytometry dataset.

You will learn

• How to prepare a dataset for training, including
  • Checking for class imbalance (Task 1.1)
  • Correcting class imbalance (Task 1.2)
  • Converting categorical data to one-hot encoding (Task 1.3)
• The basic math behind a linear classifier
• How to evaluate model performance (Tasks 2.1 - 2.3)

Part B: Random Forest Classifier

We will learn about Random Forest Classifiers and use scikit-learn to train one on our dataset.

You will learn

• How to use scikit-learn's model objects
• How to perform a hyperparameter search to optimize model performance (Tasks 3.1 - 4.1)

Part C: Feature Engineering

We will explore image filters and see if they can improve the performance of either our linear or random forest classifer.

You will learn

• How to use scikit-image's filter modules (Task 5.1)
• How to reuse what we have done so far to build a new dataset on filtered images and train your own model on the new dataset (Task 5.2)