Alpha (machine learning)

From Wikipedia, the free encyclopedia

This article is orphaned as few or no other articles link to it. Please help introduce links in articles on related topics. (September 2006)

Alpha (α) is a parameter used in artificial intelligence to specify how much weight will be given to new information. An alpha value might be expressed in the range between 0 and 1. A value of 0 means that the agent completely ignores any new information, while a value of 1 means it only remembers the latest piece of information.

A typical equation detailing the weighting role of alpha might be:

new_policy = (historical_data * (1 - α)) + (new_data * α)

It is used in Q-learning.