FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4308191114> ?p ?o ?g. }

• W4308191114 endingPage "230" @default.
• W4308191114 startingPage "207" @default.
• W4308191114 abstract "Question answering (QA) in natural language based on text corpus needs human intelligence. In this work, a system is proposed for automatic QA in natural language based on the large unstructured text corpus repository. Conventionally, computer system is more proficient in retrieving from a large data set or repository by matching keywords. Web search runs on the basic principle of keyword matching and the popular search engines improve the ranking of list of URLs collected by the crawler using page-rank algorithm, user's web cache, location, preference, etc. The principle of the QA system is different from those syntactic search algorithms. Here, actual answer is expected irrespective of number of matching terms in between question and answer. QA system built in English and other resourceful languages handles the challenge with machine readable dictionary, WordNet, ontology etc. Absence of annotated text, incomplete WordNet and ontology are the main challenges toward building QA system in the low resource language like Bengali. In the present work, supervised methods of learning algorithms are used to build Bengali QA system. Bengali literature in the form of text corpus, which was created in Technology Development of Indian Languages (TDIL) project, works as the repository of the proposed system. Four classification algorithms, ANN, SVM, Naïve Bayes, and decision tree, are used as supervised learning methods for question classification. Word2Vec algorithm is used for measuring similarity during answer retrieval. The system has achieved 95.88% accuracy in questions classification. Coarse-grain accuracy in retrieval of the answer is 98.77%. Fine-grain accuracy in retrieval of the answer is 93.25%. If the answer returned by the system is the accurate exact answer, then the result is treated as fine grain accurate; whereas if the answer returned by the system contains the exact answer, then the result is treated as coarse grain accurate. Majority of the QA system, which are built in English or other popular languages, work with a structured relational database as repository and give answer of multiple-choice type questions. Novelty of this work is in building the complete natural language-based QA system with high accuracy in low resource language, with a flat file system as repository." @default.
• W4308191114 created "2022-11-09" @default.
• W4308191114 creator A5023353760 @default.
• W4308191114 creator A5029845410 @default.
• W4308191114 date "2022-11-04" @default.
• W4308191114 modified "2023-09-26" @default.
• W4308191114 title "Question Answering System Using Deep Learning in the Low Resource Language Bengali" @default.
• W4308191114 cites W2461496554 @default.
• W4308191114 cites W2473715656 @default.
• W4308191114 cites W2477575531 @default.
• W4308191114 cites W2498816759 @default.
• W4308191114 cites W2514298988 @default.
• W4308191114 cites W2528947955 @default.
• W4308191114 cites W2565531090 @default.
• W4308191114 cites W2735189895 @default.
• W4308191114 cites W2749078814 @default.
• W4308191114 cites W2768833746 @default.
• W4308191114 cites W2804669300 @default.
• W4308191114 cites W2804989445 @default.
• W4308191114 cites W2902677179 @default.
• W4308191114 cites W2939382998 @default.
• W4308191114 cites W2990069357 @default.
• W4308191114 cites W2996659147 @default.
• W4308191114 cites W3005626884 @default.
• W4308191114 cites W3014592055 @default.
• W4308191114 cites W3101968993 @default.
• W4308191114 cites W3102887138 @default.
• W4308191114 cites W3212666329 @default.
• W4308191114 doi "https://doi.org/10.1002/9781119857686.ch10" @default.
• W4308191114 hasPublicationYear "2022" @default.
• W4308191114 type Work @default.
• W4308191114 citedByCount "0" @default.
• W4308191114 crossrefType "other" @default.
• W4308191114 hasAuthorship W4308191114A5023353760 @default.
• W4308191114 hasAuthorship W4308191114A5029845410 @default.
• W4308191114 hasConcept C111472728 @default.
• W4308191114 hasConcept C12267149 @default.
• W4308191114 hasConcept C136764020 @default.
• W4308191114 hasConcept C13743948 @default.
• W4308191114 hasConcept C138885662 @default.
• W4308191114 hasConcept C154945302 @default.
• W4308191114 hasConcept C157659113 @default.
• W4308191114 hasConcept C19235068 @default.
• W4308191114 hasConcept C195324797 @default.
• W4308191114 hasConcept C204321447 @default.
• W4308191114 hasConcept C23123220 @default.
• W4308191114 hasConcept C25810664 @default.
• W4308191114 hasConcept C41008148 @default.
• W4308191114 hasConcept C52001869 @default.
• W4308191114 hasConceptScore W4308191114C111472728 @default.
• W4308191114 hasConceptScore W4308191114C12267149 @default.
• W4308191114 hasConceptScore W4308191114C136764020 @default.
• W4308191114 hasConceptScore W4308191114C13743948 @default.
• W4308191114 hasConceptScore W4308191114C138885662 @default.
• W4308191114 hasConceptScore W4308191114C154945302 @default.
• W4308191114 hasConceptScore W4308191114C157659113 @default.
• W4308191114 hasConceptScore W4308191114C19235068 @default.
• W4308191114 hasConceptScore W4308191114C195324797 @default.
• W4308191114 hasConceptScore W4308191114C204321447 @default.
• W4308191114 hasConceptScore W4308191114C23123220 @default.
• W4308191114 hasConceptScore W4308191114C25810664 @default.
• W4308191114 hasConceptScore W4308191114C41008148 @default.
• W4308191114 hasConceptScore W4308191114C52001869 @default.
• W4308191114 hasLocation W43081911141 @default.
• W4308191114 hasOpenAccess W4308191114 @default.
• W4308191114 hasPrimaryLocation W43081911141 @default.
• W4308191114 hasRelatedWork W2069398544 @default.
• W4308191114 hasRelatedWork W2099585516 @default.
• W4308191114 hasRelatedWork W2326857978 @default.
• W4308191114 hasRelatedWork W2770028682 @default.
• W4308191114 hasRelatedWork W2943173474 @default.
• W4308191114 hasRelatedWork W3100537340 @default.
• W4308191114 hasRelatedWork W3105168542 @default.
• W4308191114 hasRelatedWork W3192589309 @default.
• W4308191114 hasRelatedWork W3204019825 @default.
• W4308191114 hasRelatedWork W2182051717 @default.
• W4308191114 isParatext "false" @default.
• W4308191114 isRetracted "false" @default.
• W4308191114 workType "other" @default.