Machine LearningTutorial | Part 3 Resampling 1: Random Undersampling | Rohit Ghosh | GreyAtom

preview_player
Показать описание
In the previous videos of the series, Rohit has talked about how to evaluate data that has been processed for quality. However, that is one way to deal with the challenges of machine learning. The second way to deal with imbalanced data is to change the data, or resampling techniques.

Assuming that there are two classes of data with differing numbers of datapoints - a classic imbalance problem. There are two options before you: either you decrease the dataset with more points to the level of the other one; or increase the dataset with fewer points to match the other one.

The first way is known as undersampling, and the second is oversampling.

Undersampling techniques include:
- Random undersampling
- Cluster centroids
- Tomek links

Watch:

#greyatom #datascience101 #challengesofmachinelearning #tutorial #machinelearning #imbalanceddata #resampling #undersampling

To stay updated with the latest trends, visit:
Рекомендации по теме