Skip to content

A leaf node for your machine learning journey, from scratch to practical applications...

License

Notifications You must be signed in to change notification settings

realamirhe/leaf-node

Repository files navigation

MLLN

logo

Machine Learning Leaf Node

An optimal start node for beginning your machine learning journey

📁 algorithms

📁 applications

🆕 🔥 SMLFDL - 📄article

SVMs multi-class loss feedback based discriminative dictionary learning for image classification

SMLFDL integrates dictionary learning and support vector machines training into a unified learning framework by looping the designed multi-class loss term, which is inspired by the feedback mechanism in cybernetics.

analysis has been done on scene-15 dataset.
Feature vectors has been prepared by four-level spatial pyramid, dense DAISY feature description followed by PCA.
As article proposed SMLFDL are faster in predictions and converge in lower epochs.
code for features will be added soon.