FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W604586230> ?p ?o ?g. }

• W604586230 abstract "The Bayesian network (BN) formalism is a powerful representation forencoding domains characterised by uncertainty. However, before itcan be used it must first be constructed, which is a major challengefor any real-life problem. There are two broad approaches, namelythe hand-crafted approach, which relies on a human expert, and thedata-driven approach, which relies on data. The former approach isuseful, however issues such as human bias can introduce errors intothe model. We have conducted a literature review of theexpert-driven approach, and we have cherry-picked a number of commonmethods, and engineered a framework to assist non-BN experts withexpert-driven construction of BNs. The latter construction approachuses algorithms to construct the model from a data set. However,construction from data is provably NP-hard.To solve this problem, approximate, heuristic algorithms have beenproposed; in particular, algorithms that assume an order between thenodes, therefore reducing the search space. However, traditionally,this approach relies on an expert providing the order among thevariables – an expert may not always be available, or may be unable toprovide the order. Nevertheless, if a good order is available, theseorder-based algorithms have demonstrated good performance. Morerecent approaches attempt to ``learn'' a good order then use theorder-based algorithm to discover the structure. To eliminate theneed for order information during construction, we propose a searchin the entire space of Bayesian network structures  – we present anovel approach for carrying out this task, and we demonstrate itsperformance against existing algorithms that search in the entirespace and the space of orders.Finally, we employ the hand-crafting framework to construct modelsfor the task of diagnosis in a ``real-life'' medical domain,dementia diagnosis. We collect real dementia data from clinicalpractice, and we apply the data-driven algorithms developed toassess the concordance between the reference models developed byhand and the models derived from real clinical data." @default.
• W604586230 created "2016-06-24" @default.
• W604586230 creator A5073179130 @default.
• W604586230 date "2008-07-01" @default.
• W604586230 modified "2023-09-24" @default.
• W604586230 title "Bayesian belief networks for dementia diagnosis and other applications: a comparison of hand-crafting and construction using a novel data driven technique" @default.
• W604586230 cites W111349384 @default.
• W604586230 cites W118859620 @default.
• W604586230 cites W1492767570 @default.
• W604586230 cites W1494546415 @default.
• W604586230 cites W1504402099 @default.
• W604586230 cites W1505477995 @default.
• W604586230 cites W1510802921 @default.
• W604586230 cites W1511622053 @default.
• W604586230 cites W1515265418 @default.
• W604586230 cites W1518076737 @default.
• W604586230 cites W1518177337 @default.
• W604586230 cites W1519871408 @default.
• W604586230 cites W1521567963 @default.
• W604586230 cites W1523680690 @default.
• W604586230 cites W1524326598 @default.
• W604586230 cites W1537120672 @default.
• W604586230 cites W1558174409 @default.
• W604586230 cites W1559294607 @default.
• W604586230 cites W1560253955 @default.
• W604586230 cites W1560883996 @default.
• W604586230 cites W1563348628 @default.
• W604586230 cites W1565206817 @default.
• W604586230 cites W1568153837 @default.
• W604586230 cites W1577668191 @default.
• W604586230 cites W1586003574 @default.
• W604586230 cites W158727920 @default.
• W604586230 cites W1595498733 @default.
• W604586230 cites W1606458208 @default.
• W604586230 cites W1659842140 @default.
• W604586230 cites W166035671 @default.
• W604586230 cites W1668763905 @default.
• W604586230 cites W1669437150 @default.
• W604586230 cites W1730193065 @default.
• W604586230 cites W1755360231 @default.
• W604586230 cites W1790876067 @default.
• W604586230 cites W1817966645 @default.
• W604586230 cites W1820051232 @default.
• W604586230 cites W1847168837 @default.
• W604586230 cites W194882608 @default.
• W604586230 cites W1959606346 @default.
• W604586230 cites W1964735403 @default.
• W604586230 cites W1969585839 @default.
• W604586230 cites W1970345407 @default.
• W604586230 cites W1972978214 @default.
• W604586230 cites W1973107043 @default.
• W604586230 cites W1975062332 @default.
• W604586230 cites W1980888617 @default.
• W604586230 cites W198094292 @default.
• W604586230 cites W1983399068 @default.
• W604586230 cites W1985174640 @default.
• W604586230 cites W1986511578 @default.
• W604586230 cites W1989908024 @default.
• W604586230 cites W1989926363 @default.
• W604586230 cites W1990194281 @default.
• W604586230 cites W1990411923 @default.
• W604586230 cites W1995307465 @default.
• W604586230 cites W1996034668 @default.
• W604586230 cites W1998600012 @default.
• W604586230 cites W1999299399 @default.
• W604586230 cites W2003403599 @default.
• W604586230 cites W2004920427 @default.
• W604586230 cites W2008906462 @default.
• W604586230 cites W2009075080 @default.
• W604586230 cites W2010801381 @default.
• W604586230 cites W2012138753 @default.
• W604586230 cites W2015686022 @default.
• W604586230 cites W201777379 @default.
• W604586230 cites W2017987083 @default.
• W604586230 cites W2018608850 @default.
• W604586230 cites W2019057855 @default.
• W604586230 cites W2021071494 @default.
• W604586230 cites W2021453862 @default.
• W604586230 cites W2022200205 @default.
• W604586230 cites W2022703447 @default.
• W604586230 cites W2023461130 @default.
• W604586230 cites W2028076981 @default.
• W604586230 cites W2029800614 @default.
• W604586230 cites W2031180623 @default.
• W604586230 cites W2031709382 @default.
• W604586230 cites W2032072348 @default.
• W604586230 cites W2034622004 @default.
• W604586230 cites W2034962152 @default.
• W604586230 cites W2036559064 @default.
• W604586230 cites W2043815141 @default.
• W604586230 cites W2044980147 @default.
• W604586230 cites W2046690910 @default.
• W604586230 cites W2049053063 @default.
• W604586230 cites W2051753984 @default.
• W604586230 cites W2051996699 @default.
• W604586230 cites W2055037429 @default.
• W604586230 cites W2056309070 @default.
• W604586230 cites W2056880208 @default.
• W604586230 cites W2057679568 @default.
• W604586230 cites W2058978992 @default.

• next