- Three Nested Adaptive Loops
- For every iteration of outer Loop I, there are many iterations of intermediate Loop II.
- For every iteration of intermediate Loop II, there are many iterations of inner Loop III.
|
- Decisions for Loop III
- Is discrimination acceptable for final classification?
- Yes: Stop.
- No: Is discrimination increasing fast enough to continue processing with current candidate feature set?
- Yes: Continue with Weight Adjustment.
- No: Go to Loop II.
|
- Decisions for Loop II
- Is progress rate acceptable for trying different frequency feature sets based on the current form of amplitude function, f(
)?
- Yes: Continue with frequency feature adjustment.
- No: Go to Loop I.
|
- Decisions for Loop I
- Is progress rate acceptable with the current method of generating candidate exponential forms of f(
)?
- Yes: Continue with adjustment of the amplitude and exponential coefficients of f(
).
- No: Stop and configure another method for generating candidate forms of f(
).
|