|
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.
42
d) The network tested with words Speaker1
Actual language Czech English Hungarian ∑
All words 153
Precisely identified words 124
Efficiency 74,51% 90,20% 78,43% 81,05%
Imprecisely identified words
(instead the actual language)
as
Hungarian
as
English
as
Hungarian
as
Czech
as
English
as
Czech
∑
8 29
Ratio 15,69% 9,80% 7,84% 1,96% 11,76% 9,80% 18,95%
Table 8. The results presented hereinafter
were obtained average training and tests per train (with test_3, test_4
and test_5 folder) total number tests.13 The results the NNV2 network tested with words Speaker2
f) 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.
Actual language Czech English Hungarian ∑
All words 153
Precisely identified words 70
Efficiency 33,33% 45,10% 58,82% 45,75%
Imprecisely identified words
(instead the actual language)
as
Hungarian
as
English
as
Hungarian
as
Czech
as
English
as
Czech
∑
18 83
Ratio 35,29% 31,37% 43,14% 11,76% 37,25% 3,92% 54,25%
Table 8.12 The results the NNV2 network tested with words Speaker1
e) The network tested with words Speaker2
Actual language Czech English Hungarian ∑
All words 153
Precisely identified words 116
Efficiency 70,59% 70,59% 86,27% 75,82%
Imprecisely identified words
(instead the actual language)
as
Hungarian
as
English
as
Hungarian
as
Czech
as
English
as
Czech
∑
6 37
Ratio 11,76% 17,65% 23,53% 5,88% 5,88% 7,84% 24,18%
Table 8.14 The results the NNV2 network tested with words Speaker3
