izpis_h1_title_alt

Samodejno reševanje slovenskih križank s pomočjo spleta
ID TESOVNIK, DEJAN (Author), ID Kononenko, Igor (Mentor) More about this mentor... This link opens in a new window

.pdfPDF - Presentation file, Download (1,10 MB)
MD5: 9D00387F94D402FFD3D2D3AD04646AB5
PID: 20.500.12556/rul/52642cbb-7699-4342-bf67-963e67c06b04

Abstract
Cilj diplomske naloge je bil izdelati program, ki samostojno rešuje slovenske križanke s pomočjo ugankarskega slovarja in spletnega iskalnika. Za izdelavo programa je bila uporabljena baza 367 križank in ugankarski slovar s 104.132 gesli, slovar spletnih odgovorov pa je bil izdelan s pomočjo spletnega iskalnika. Zaradi ogromnega števila spletnih odgovorov bi bilo preiskovanje celotnega prostora prepotratno. Reševanje križanke se zato izvede v dveh ločenih fazah. Program v prvi fazi delno reši križanko le s pomočjo ugankarskega slovarja. V tej fazi program najde najprimerjnejše odgovore za vsak namig v križanki in z njimi izpolni križanko. V drugi fazi program vključi v reševanje tudi spletne odgovore ter tako pride do končne rešitve. Program je bil razvit s pomočjo učne množice 37 križank, uspešnost programa pa je bila preizkušena na testni množici 330 križank.

Language:Slovenian
Keywords:križanke, preiskovalni algoritmi, avtomatsko reševanje, slovar spletnih odgovorov
Work type:Bachelor thesis/paper
Organization:FRI - Faculty of Computer and Information Science
Year:2018
PID:20.500.12556/RUL-99546 This link opens in a new window
Publication date in RUL:31.01.2018
Views:6278
Downloads:539
Metadata:XML DC-XML DC-RDF
:
Copy citation
Share:Bookmark and Share

Secondary language

Language:English
Title:Automated Slovene crosswords solving with the help of the internet
Abstract:
The goal of the thesis was to create a program, that automatically solves Slovene crosswords with the help of puzzle dictionary and web search. For the purpose of creating the program, we used a database of 367 crosswords and puzzle dictionary with 104.123 clue answers. We also created a list of web answers for all the crosswords in the database. Because of the enormous amount of the web answers searching the whole solution space would not be feasible. That is why the program solves each crossword in two separate phases. In the first phase, a crossword gets partially solved with the help of the puzzle dictionary, which creates a list of most probable solution candidates for each clue. In the second phase, a crossword is solved with puzzle dictionary and web answers combined. For the purpose of building the program, we used a set of 37 training crosswords. The final results were obtained from a set of 330 testing crosswords.

Keywords:crosswords, search algorithms, automatic solving, web search dictionary

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Back