Handling Imbalanced Data Classification Problems

In this 2-hour long project-based course on handling imbalanced data classification problems, you will learn to understand the business problem related we are trying to solve and and understand the dataset. You will also learn how to select best evaluation metric for imbalanced datasets and data resampling techniques like undersampling, oversampling and SMOTE before we use them for model building process. At the end of the course you will understand and learn how to implement ROC curve and adjust probability threshold to improve selected evaluation metric of the model. Note: This course works best for learners who are based in the North America region. We’re currently working on providing the same experience in other regions.

• Project Overview
• Welcome to this project-based course on Handling Imbalanced Data Classification Problems. In this project, you will learn how to apply various data resampling techniques like undersampling, oversampling, SMOTE on the imbalanced datasets and be able to build a classifier to identify or predict the minority class samples. By the end of this 2-hour long project, you will be able to understand what imbalanced datasets are, what are the evaluation metrics that we should consider while building imbalanced data classification models. You will build predictive models after resampling to balance the classes of target variables and use ROC curve to adjust probability threshold which will help you to improve the evaluation metric of your choice.