DELTA RULE
Charles François (2004). DELTA RULE, International Encyclopedia of Systems and Cybernetics, 2(1): 848.
| Collection | International Encyclopedia of Systems and Cybernetics |
|---|---|
| Year | 2004 |
| Vol. (num.) | 2(1) |
| ID | ◀ 848 ▶ |
| Object type | Methodology or model |
In parallel processing networks , “A learning rule that utilizes the discrepancy between the desired and actual output of each output unit to change the weigths feeding into it” (W. BECHTEL & A. ABRAHAMSON, 1991, p. 306).
- “It is also known as the WIDROW-HOFF rule and the least mean squares rule” (p.306).
This amounts to an adaptive process based on feedbacks and is the basic tool for the progressive betterment of the efficiency of the learning network .
See also
Learning (Back propagation)