Imbalanced classification are those prediction tasks where the distribution of examples across class labels is not equal.<\/p>\n
Most imbalanced classification examples focus on binary classification tasks, yet many of the tools and techniques for imbalanced classification also directly support multi-class classification problems.<\/p>\n
In this tutorial, you will discover how to use the tools of imbalanced classification with a multi-class dataset.<\/p>\n
After completing this tutorial, you will know:<\/p>\n
- \n
- About the glass identification standard imbalanced multi-class prediction problem.<\/li>\n
- How to use SMOTE oversampling for imbalanced multi-class classification.<\/li>\n
- How to use cost-sensitive learning for imbalanced multi-class classification.<\/li>\n<\/ul>\n
Discover SMOTE, one-class classification, cost-sensitive learning, threshold moving, and much more in my new book<\/a>, with 30 step-by-step tutorials and full Python source code.<\/p>\n
Let’s get started.<\/p>\n
![\"Multi-Class](\"https:\/\/machinelearningmastery.com\/wp-content\/uploads\/2020\/08\/Multi-Class-Imbalanced-Classification.jpg\")
<\/p>\nMulti-Class Imbalanced Classification
Photo by istolethetv<\/a>, some rights reserved.<\/p>\n<\/div>\nTutorial Overview<\/h2>\n
This tutorial is divided into three parts; they are:<\/p>\n
- \n
- Glass Multi-Class Classification Dataset<\/li>\n
- SMOTE Oversampling for Multi-Class Classification<\/li>\n
- Cost-Sensitive Learning for Multi-Class Classification<\/li>\n<\/ol>\n
Glass Multi-Class Classification Dataset<\/h2>\n
In this tutorial, we will focus on the standard imbalanced multi-class classification problem referred to as “Glass Identification<\/strong>” or simply “glass<\/em>.”<\/p>\n
The dataset describes the chemical properties of glass and involves classifying samples of glass using their chemical properties as one of six classes. The dataset was credited to Vina Spiehler<\/a> in 1987.<\/p>\n
Ignoring the sample identification number, there are nine input variables that summarize the properties of the glass dataset; they are:<\/p>\n
- \n
- RI: Refractive Index<\/li>\n
- Na: Sodium<\/li>\n
- Mg: Magnesium<\/li>\n
- Al: Aluminum<\/li>\n
- Si: Silicon<\/li>\n
- K: Potassium<\/li>\n
- Ca: Calcium<\/li>\n
- Ba: Barium<\/li>\n
- Fe: Iron<\/li>\n<\/ul>\n
The chemical compositions are measured as the weight percent in corresponding oxide.<\/p>\n
There are seven types of glass listed; they are:<\/p>\n
- \n
- Class 1: building windows (float processed)<\/li>\n
- Class 2: building windows (non-float processed)<\/li>\n
- Class 3: vehicle windows (float processed)<\/li>\n
- Class 4: vehicle windows (non-float processed)<\/li>\n
- Class 5: containers<\/li>\n
- Class 6: tableware<\/li>\n
- Class 7: headlamps<\/li>\n<\/ul>\n
Float glass<\/a> refers to the process used to make the glass.<\/p>\n
There are 214 observations in the dataset and the number of observations in each class is imbalanced. Note that there are no examples for class 4 (non-float processed vehicle windows) in the dataset.<\/p>\n
- \n
- Class 1: 70 examples<\/li>\n
- Class 2: 76 examples<\/li>\n
- Class 3: 17 examples<\/li>\n
- Class 4: 0 examples<\/li>\n
- Class 5: 13 examples<\/li>\n
- Class 6: 9 examples<\/li>\n
- Class 7: 29 examples<\/li>\n<\/ul>\n
Although there are minority classes, all classes are equally important in this prediction problem.<\/p>\n
The dataset can be divided into window glass (classes 1-4) and non-window glass (classes 5-7). There are 163 examples of window glass and 51 examples of non-window glass.<\/p>\n
- \n
- Window Glass: 163 examples<\/li>\n
- Non-Window Glass: 51 examples<\/li>\n<\/ul>\n
Another division of the observations would be between float processed glass and non-float processed glass, in the case of window glass only. This division is more balanced.<\/p>\n
- \n
- Float Glass: 87 examples<\/li>\n
- Non-Float Glass: 76 examples<\/li>\n<\/ul>\n
You can learn more about the dataset here:<\/p>\n
- \n
- Glass Dataset (glass.csv)<\/a><\/li>\n
- Glass Dataset Description (glass.names)<\/a><\/li>\n<\/ul>\n
No need to download the dataset; we will download it automatically as part of the worked examples.<\/p>\n
Below is a sample of the first few rows of the data.<\/p>\n
1.52101,13.64,4.49,1.10,71.78,0.06,8.75,0.00,0.00,1\r\n1.51761,13.89,3.60,1.36,72.73,0.48,7.83,0.00,0.00,1\r\n1.51618,13.53,3.55,1.54,72.99,0.39,7.78,0.00,0.00,1\r\n1.51766,13.21,3.69,1.29,72.61,0.57,8.22,0.00,0.00,1\r\n1.51742,13.27,3.62,1.24,73.08,0.55,8.07,0.00,0.00,1\r\n...<\/pre>\n
We can see that all inputs are numeric and the target variable in the final column is the integer encoded class label.<\/p>\n
You can learn more about how to work through this dataset as part of a project in the tutorial:<\/p>\n
- Glass Dataset Description (glass.names)<\/a><\/li>\n<\/ul>\n
- Glass Dataset (glass.csv)<\/a><\/li>\n