[HIVEMALL-222] Introduce Gradient Clipping to avoid exploding gradient to General Classifier/Regressor

• Type: Improvement
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 0.5.0
• Fix Version/s: 0.5.2
• Assignee: Makoto Yui
• Reporter: Makoto Yui
• Due: 31/Oct/18
• Created: 24/Oct/18 05:50
• Updated: 24/Oct/18 08:28
• Resolved: 24/Oct/18 08:28

Description

Gradient Clipping is useful for avoiding exploding gradients
https://github.com/scikit-learn/scikit-learn/blob/0fc7ce6bb8bb6b5b98c66ad2c0a009753def945a/sklearn/linear_model/sgd_fast.pyx#L701

Details of exploding gradients can be found in this link.

So, implement it for General Classifier/Regressor

Issue Links

• links to: GitHub Pull Request #169