|
Kategorie: Diplomové, bakalářské práce |
Tento dokument chci!
Práce popisuje základy principu funkčnosti neuronů a vytvoření umělých neuronových sítí. Je zde důkladně popsána struktura a funkce neuronů a ukázán nejpoužívanější algoritmus pro učení neuronů. Základy fuzzy logiky, včetně jejich výhod a nevýhod, jsou rovněž prezentovány. Detailněji je popsán algoritmus zpětného šíření chyb a adaptivní neuro-fuzzy inferenční systém. Tyto techniky poskytují efektivní způsoby učení neuronových sítí.
41
b) The network tested with words Speaker2
Actual language Czech English Hungarian ∑
All words 153
Precisely identified words 100
Efficiency 68,63% 62,75% 64,71% 65,36%
Imprecisely identified words
(instead the actual language)
as
Hungarian
as
English
as
Hungarian
as
Czech
as
English
as
Czech
∑
9 53
Ratio 17,65% 13,73% 17,65% 19,61% 17,65% 17,65% 34,64%
Table 8.11 The results the NNV1 network tested with words Speaker3
For the second NNV network (NNV2), the same simulations were run with the
analysis parameters set params.
.10 The results the NNV1 network tested with words Speaker2
c) The network tested with words Speaker3
For this simulation, the network was trained with the words Speaker1 and
Speaker2 and then tested with words Speaker3.m follows:
framestep=30; %ms
framelen=35; %ms
melfilerbankcount=10;
With this setting, one simulation took approximately 390 seconds. The results presented hereinafter
were obtained average training and tests per train (with test_3, test_4
and test_5 folder) total number tests.
Actual language Czech English Hungarian ∑
All words 153
Precisely identified words 66
Efficiency 37,25% 31,37% 60,78% 43,14%
Imprecisely identified words
(instead the actual language)
as
Hungarian
as
English
as
Hungarian
as
Czech
as
English
as
Czech
∑
19 87
Ratio 37,25% 25,49% 62,75% 5,88% 35,29% 3,92% 56,86%
Table 8
