Experimenting with computing the expected event probability on each AdaBoostian step.
A nicer printing of probability branches in ex26, in a column rather than a row.
in ex26 added support of the arbitrary training confidences, with interpolation
Looks like ex26 works. Added a couple of input examples for it with confidence 0.5.
Fixed the crash in ex26 (but the result iis not correct yet).
Work in progress on ex26, computing the adaboostian table with multiple preceding variants. Doesn't work right yet.
ex26 now works backwards-compatibly with ex24.
Work in progress on supporting partial confidences in Adaboostian version.