December 2018
Intermediate to advanced
318 pages
8h 28m
English
Synthetic Minority Oversampling Technique (SMOTE) is a technique where synthetic data is generated by taking a subset of the data from the minority classes. However, none of the data is a replica of that in the minority class, thus overfitting is easily avoided. The synthetic data is added to the original dataset. This combined dataset is used to classify data. The good thing about this sampling is that there is no loss of information during the entire process.
Read now
Unlock full access