Pregled bibliografske jedinice broj: 931711
Detecting Non-covered Questions in Frequently Asked Questions Collections
Detecting Non-covered Questions in Frequently Asked Questions Collections // Proceedings of the International Conference on Applications of Natural Language to Information Systems. / Frasincar, Flavius ; Ittoo, Ashwin ; Nguyen, Le Minh ; Métais, Elisabeth (ur.).
Liege: Springer, 2017. str. 387-390 doi:10.1007/978-3-319-59569-6_47 (poster, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)
CROSBI ID: 931711 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Detecting Non-covered Questions in Frequently Asked Questions Collections
Autori
Karan, Mladen ; Šnajder, Jan
Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni
Izvornik
Proceedings of the International Conference on Applications of Natural Language to Information Systems.
/ Frasincar, Flavius ; Ittoo, Ashwin ; Nguyen, Le Minh ; Métais, Elisabeth - Liege : Springer, 2017, 387-390
ISBN
978-3-319-59569-6
Skup
International Conference on Applications of Natural Language to Information Systems.
Mjesto i datum
Liège, Belgija, 21.06.2017. - 23.06.2017
Vrsta sudjelovanja
Poster
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
FAQ retrieval, Novelty detection, Question answering
Sažetak
Frequently asked questions (FAQ) collections are a popular and effective way of representing information, and FAQ retrieval systems provide a natural-language interface to such collections. An important aspect of efficient and trustworthy FAQ retrieval is to maintain a low fall-out rate by detecting non-covered questions. In this paper we address the task of detecting non-covered questions. We experiment with threshold-based methods as well as unsupervised one-class and supervised binary classifiers, considering tf-idf and word embeddings text representations. Experiments, carried out on a domain-specific FAQ collection, indicate that a cluster-based model with query paraphrases outperforms threshold-based, one-class, and binary classifiers.
Izvorni jezik
Engleski
Znanstvena područja
Računarstvo
POVEZANOST RADA
Projekti:
036-1300646-1986 - Otkrivanje znanja u tekstnim podacima (Dalbelo-Bašić, Bojana, MZO ) ( CroRIS)
Ustanove:
Fakultet elektrotehnike i računarstva, Zagreb