Imbalanced Classification Master Class in Python

A Step-by-Step Guide to Handling Real-World Class Imbalance in Machine Learning

Welcome to Imbalanced Classification Master Class in Python.

What you’ll learn

  • How to use data sampling algorithms like SMOTE to transform the training dataset for an imbalanced dataset when fitting a range of machine learning models.
  • How algorithms from the field of cost-sensitive learning can be used for imbalanced classification.
  • How to use modified versions of standard algorithms like SVM and decision trees to take the class weighting into account.
  • How to tune the threshold when interpreting predicted probabilities as class labels.
  • How to calibrate probabilities predicted by nonlinear algorithms that are not fit using a probabilistic framework.
  • How to use algorithms from the field of outlier detection and anomaly detection for imbalanced classification.
  • How to use modified ensemble algorithms that have been modified to take the class distribution into account during training.
  • How to systematically work through an imbalanced classification predictive modeling project.

Course Content

  • Introduction –> 7 lectures • 12min.
  • Understanding Class Imbalance –> 7 lectures • 17min.
  • Model Evaluation –> 28 lectures • 1hr 1min.
  • Data Sampling –> 20 lectures • 42min.
  • Cost-Sensitive Learning –> 19 lectures • 37min.
  • Projects –> 7 lectures • 17min.

Imbalanced Classification Master Class in Python

Requirements

  • You’ll need a solid foundation in machine learning.
  • You’ll need a solid background in Python.
  • A familiarity with classification problems would be ideal.

Welcome to Imbalanced Classification Master Class in Python.

Classification predictive modeling is the task of assigning a label to an example. Imbalanced classification is those classification tasks where the distribution of examples across the classes is not equal. Typically the class distribution is severely skewed so that for each example in the minority class, there may be one hundred or even one thousand examples in the majority class. Practical imbalanced classification requires the use of a suite of specialized techniques, data preparation techniques, learning algorithms, and performance metrics.

Let’s discuss what you’ll learn in this course.

  • The challenge and intuitions for imbalanced classification datasets.
  • How to choose an appropriate performance metric for evaluating models for imbalanced classification.
  • How to appropriately stratify an imbalanced dataset when splitting into train and test sets and when using k-fold cross-validation.
  • How to use data sampling algorithms like SMOTE to transform the training dataset for an imbalanced dataset when fitting a range of standard machine learning models.
  • How algorithms from the field of cost-sensitive learning can be used for imbalanced classification.
  • How to use modified versions of standard algorithms like SVM and decision trees to take the class weighting into account.
  • How to tune the threshold when interpreting predicted probabilities as class labels.
  • How to calibrate probabilities predicted by nonlinear algorithms that are not fit using a probabilistic framework.
  • How to use algorithms from the field of outlier detection and anomaly detection for imbalanced classification.
  • How to use modified ensemble algorithms that have been modified to take the class distribution into account during training.
  • How to systematically work through an imbalanced classification predictive modeling project.

This course was created to be completed linearly, from start to finish. That being said, if you know the basics and need help with a specific method or type of problem, then you can flip straight to that section and get started. This course was designed for you to completed on your laptop or desktop, on the screen, not on a tablet.

My hope is that you have the course open right next to your editor and run the examples as you read about them. This course is not intended to be completed passively or be placed in a folder as a reference text. It is a playbook, a workbook, and a guidebook intended for you to learn by doing and then apply your new understanding with working Python examples. To get the most out of the course, I would recommend playing with the examples in each tutorial. Extend them, break them, then fix them.

Thanks for you interest in Imbalanced Classification Master Class in Python.

Now let’s get started!