2500+ classes across the 15 features for classification model. | Coursera Community

2500+ classes across the 15 features for classification model.

  • 5 April 2021
  • 1 reply
  • 6 views
P
Badge +1

How do you reduce the number of classes for scikit-learn models? Which steps will you follow for feature enginnering?

1 reply

devp
Userlevel 4
Badge +7

@parthasen Assuming you are asking to identify truly different classes,first use the Elbow method to identify the number of truly different classes. Then use unsupervised Kmeans clustering to cluster the data points.

Then, I would attribute the class value [lesser than 2500 now] to each data point.

Reply


coursera footer

Coursera provides universal access to the world’s best education, partnering with top universities and organizations to offer courses online.

© 2020 Coursera Inc. All rights reserved.

Download on the App Store Get it on Google Play