ALL libraries (COBIB.SI union bibliographic/catalogue database)
-
Podatkovno vodeni pristop napovedovanja prozodičnih mej z arhitekturami nevronskih mrež [Elektronski vir]Stergar, Janez, 1968-Podatkovno vodeni pristopi (PVP) so postopki, s katerih učimo avtoasociativne strukture z izbranimi parametri iz obsežnih podatkovnih baz. PVP predstavljajo trend na področjih govornih tehnologij ... (adaptaciji sistemov in aplikacij). V kategorijo pomembnih parametrov, ki prispevajo k razumljivosti, predvsem pa k naravnosti sintetiziranega govora, uvrščamo prozodične parametre. PVP omogočajo samodejno izločitev prozodičnih parametrov z analizo ustrezne besedilne zbirke, hkrati pa pogojujejo doslednost in konsistentnost pri njihovem označevanju. Uspešnost napovedovanja prozodičnih parametrov zavisi od postopka označevanja besedilne zbirke. Če smo pri napovedovanju simbolične prozodije - prozodije na nivoju besed, besednih zvez, stavkov in povedi - uspešni, je to pomemben korak k izboljšanju naravnosti sintetiziranega govora. V članku bomo predstavili postopek napovedovanja prozodičnih mej, ki smo jih izločili s samodejno/ročnim označevanjem besedilne zbirke. Preskusili smo uspešnost njihovega napovedovanja z uporabo dveh različnih arhitektur nevronskih mrež (NN). Rezultati uspešnosti napovedovanja izpostavljajo učinkovitost metode označevanja besedilne zbirke, saj smo dosegli dobre rezultate napovedovanja prozodičnih mej že s prepolovljenim obsegom besedilne zbirke (600 povedi). Z arhitekturo klasifikacije napak samodejnoasociativnih nevronskih mrež smo dosegli uspešnost napovedovanja 89,9 %.Source: Raziskovalno delo podiplomskih študentov v Sloveniji - ena znanost [Elektronski vir] : e-zbornik (Str. 415-423)Type of material - e-articlePublish date - 2003Language - slovenianCOBISS.SI-ID - 7852822
Author
Stergar, Janez, 1968-
Topics
sinteza govora |
simbolna prozodija |
prozodične meje |
nevronske mreže |
večnivojski perceptron |
samodejno-asociativne nevronske mreže |
speech synthesis |
symbolic prosody |
prosody boundaries |
neural networks |
multi layer perceptron |
autoassociator classificator architecture
Shelf entry
Permalink
- URL:
Impact factor
Access to the JCR database is permitted only to users from Slovenia. Your current IP address is not on the list of IP addresses with access permission, and authentication with the relevant AAI accout is required.
Year | Impact factor | Edition | Category | Classification | ||||
---|---|---|---|---|---|---|---|---|
JCR | SNIP | JCR | SNIP | JCR | SNIP | JCR | SNIP |
Select the library membership card:
If the library membership card is not in the list,
add a new one.
DRS, in which the journal is indexed
Database name | Field | Year |
---|
Links to authors' personal bibliographies | Links to information on researchers in the SICRIS system |
---|---|
Stergar, Janez, 1968- | 14915 |
Source: Personal bibliographies
and: SICRIS
Select pickup location:
Material pickup by post
Delivery address:
Address is missing from the member's data.
The address retrieval service is currently unavailable, please try again.
By clicking the "OK" button, you will confirm the pickup location selected above and complete the reservation process.
By clicking the "OK" button, you will confirm the above pickup location and delivery address, and complete the reservation process.
By clicking the "OK" button, you will confirm the address selected above and complete the reservation process.
Notification
Automatic login and reservation service currently not available. You can reserve the material on the Biblos portal or try again here later.
Subject headings in COBISS General List of Subject Headings
Select pickup location
The material from the parent unit is free. If the material is delivered to the pickup location from another unit, the library may charge you for this service.
Pickup location | Material status | Reservation |
---|
Reservation in progress
Please wait a moment.
Reservation was successful.
Reservation failed.
Reservation...
Membership card:
Pickup location: