FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2316106883> ?p ?o ?g. }

• W2316106883 abstract "Event Abstract Back to Event Naively successful: Naïve Bayes with and without decision trees for automated annotation of human neuroimaging abstracts Chayan Chakrabarti1, Thomas B. Jones1, Angela R. Laird2, Jiawei F. Xu1, George Luger1, Jessica A. Turner3* and Matthew D. Turner1, 3 1 University of New Mexico, Computer Science, United States 2 Florida International University, United States 3 Mind Research Network, United States Introduction. We are developing prototype tools for the automated extraction of metadata from collections of fMRI abstracts. The Cognitive Paradigm Ontology (CogPO) defines ontological relationships between terms pertaining to fMRI experiments. BrainMap is a database describing these experiments, which is annotated by human experts guided by the CogPO terms. We explored the performance of a naïve Bayes classifier (with several transformation approaches for this multi-label task) and the combination of the classifiers with decision trees. These efforts aim at improving the workflow for filtering and classification of fMRI papers into usable and relevant groups for further meta-analysis, storage, or annotation within other software frameworks. Methods. The CogPO annotations for the gold standard corpus were from seven categories: Each abstract had at least one and possibly more labels drawn from Paradigm Class (e.g., Stroop), Behavioral Domain (e.g., Attention), Stimulus Types (e.g., Faces), Stimulus Modalities (e.g. Visual), Instructions (e.g. Remember), Response Types (e.g. Button Press), and Response Modalities (e.g. Foot). Using naïve Bayes and either binary relevance or a label powerset transformation, we trained the classifiers on the gold-standard set of 247 abstracts. Performance using these classifiers was then assessed for each category. To determine whether a hierarchical approach could leverage the internal structure of these categories, we created decision trees for combining these categories of annotations. We trained multiple decision trees based on each of the terms of a given category being chosen as the root node. For annotating new samples from a test set, we determine which of the terms is the likeliest tag using the naïve Bayes results. Then we follow the appropriate decision trees to obtain tags for all the categories. Results. Our naïve Bayes classifier on annotation terms in each dimension produced surprisingly good results. The integrated F1-micro measure for the various categories and methods ranged from 0.3 to above 0.8; the proportion of abstracts labeled with the exact matching set of labels (i.e. no extra labels, and all the correct labels identified) varied as well across the categories, with a mean of 0.42. The Bayesian decision trees performed slightly better than naïve Bayes on each category alone. Incorporating expert knowledge through identifying correctly one of the category labels (e.g., starting with the correct label for Stimulus Type), improved performance on the remainder of the categories. Conclusions. We recognize that journal article abstracts will likely not provide adequate information for complete classification. The ability to identify the correct set of labels in each category was influenced by the number of possible labels (categories with more labels to choose from had worse performance than categories with fewer), and by the number of labels per instance (performance was better for abstracts with fewer labels), as well as by the dependencies captured by the decision tree hierarchies. The methods presented here are generic and can be extended to the full-text of articles (or subsections). Although our algorithms cannot completely replace a human annotator, they can serve as a valuable aid to complement the annotation process for a human expert, and may lead to reduction in time and effort. Acknowledgements This project was supported by a grant from NIH to Drs. Turner and Laird, 1R56MH097870-01 Keywords: text mining, naive bayes, annotation, fMRI, Decision Trees Conference: Neuroinformatics 2013, Stockholm, Sweden, 27 Aug - 29 Aug, 2013. Presentation Type: Poster Topic: Neuroimaging Citation: Chakrabarti C, Jones TB, Laird AR, Xu JF, Luger G, Turner JA and Turner MD (2013). Naively successful: Naïve Bayes with and without decision trees for automated annotation of human neuroimaging abstracts. Front. Neuroinform. Conference Abstract: Neuroinformatics 2013. doi: 10.3389/conf.fninf.2013.09.00094 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 08 Apr 2013; Published Online: 11 Jul 2013. * Correspondence: Dr. Jessica A Turner, Mind Research Network, Albuquerque, New Mexico, 87106, United States, jessica.turner@osumc.edu Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Chayan Chakrabarti Thomas B Jones Angela R Laird Jiawei F Xu George Luger Jessica A Turner Matthew D Turner Google Chayan Chakrabarti Thomas B Jones Angela R Laird Jiawei F Xu George Luger Jessica A Turner Matthew D Turner Google Scholar Chayan Chakrabarti Thomas B Jones Angela R Laird Jiawei F Xu George Luger Jessica A Turner Matthew D Turner PubMed Chayan Chakrabarti Thomas B Jones Angela R Laird Jiawei F Xu George Luger Jessica A Turner Matthew D Turner Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page." @default.
• W2316106883 created "2016-06-24" @default.
• W2316106883 creator A5019651157 @default.
• W2316106883 creator A5025336372 @default.
• W2316106883 creator A5046825469 @default.
• W2316106883 creator A5050133756 @default.
• W2316106883 creator A5066509392 @default.
• W2316106883 creator A5070412652 @default.
• W2316106883 creator A5085013304 @default.
• W2316106883 date "2013-01-01" @default.
• W2316106883 modified "2023-09-23" @default.
• W2316106883 title "Naively successful: Naïve Bayes with and without decision trees for automated annotation of human neuroimaging abstracts" @default.
• W2316106883 doi "https://doi.org/10.3389/conf.fninf.2013.09.00094" @default.
• W2316106883 hasPublicationYear "2013" @default.
• W2316106883 type Work @default.
• W2316106883 sameAs 2316106883 @default.
• W2316106883 citedByCount "0" @default.
• W2316106883 crossrefType "journal-article" @default.
• W2316106883 hasAuthorship W2316106883A5019651157 @default.
• W2316106883 hasAuthorship W2316106883A5025336372 @default.
• W2316106883 hasAuthorship W2316106883A5046825469 @default.
• W2316106883 hasAuthorship W2316106883A5050133756 @default.
• W2316106883 hasAuthorship W2316106883A5066509392 @default.
• W2316106883 hasAuthorship W2316106883A5070412652 @default.
• W2316106883 hasAuthorship W2316106883A5085013304 @default.
• W2316106883 hasBestOaLocation W23161068831 @default.
• W2316106883 hasConcept C107673813 @default.
• W2316106883 hasConcept C119857082 @default.
• W2316106883 hasConcept C12267149 @default.
• W2316106883 hasConcept C154945302 @default.
• W2316106883 hasConcept C204321447 @default.
• W2316106883 hasConcept C207201462 @default.
• W2316106883 hasConcept C2776321320 @default.
• W2316106883 hasConcept C41008148 @default.
• W2316106883 hasConcept C52001869 @default.
• W2316106883 hasConcept C95623464 @default.
• W2316106883 hasConceptScore W2316106883C107673813 @default.
• W2316106883 hasConceptScore W2316106883C119857082 @default.
• W2316106883 hasConceptScore W2316106883C12267149 @default.
• W2316106883 hasConceptScore W2316106883C154945302 @default.
• W2316106883 hasConceptScore W2316106883C204321447 @default.
• W2316106883 hasConceptScore W2316106883C207201462 @default.
• W2316106883 hasConceptScore W2316106883C2776321320 @default.
• W2316106883 hasConceptScore W2316106883C41008148 @default.
• W2316106883 hasConceptScore W2316106883C52001869 @default.
• W2316106883 hasConceptScore W2316106883C95623464 @default.
• W2316106883 hasLocation W23161068831 @default.
• W2316106883 hasOpenAccess W2316106883 @default.
• W2316106883 hasPrimaryLocation W23161068831 @default.
• W2316106883 hasRelatedWork W1493451373 @default.
• W2316106883 hasRelatedWork W1506481126 @default.
• W2316106883 hasRelatedWork W2095916325 @default.
• W2316106883 hasRelatedWork W2296661347 @default.
• W2316106883 hasRelatedWork W2780177025 @default.
• W2316106883 hasRelatedWork W2948288905 @default.
• W2316106883 hasRelatedWork W2997277234 @default.
• W2316106883 hasRelatedWork W4224221743 @default.
• W2316106883 hasRelatedWork W4251019512 @default.
• W2316106883 hasRelatedWork W4288333917 @default.
• W2316106883 hasVolume "7" @default.
• W2316106883 isParatext "false" @default.
• W2316106883 isRetracted "false" @default.
• W2316106883 magId "2316106883" @default.
• W2316106883 workType "article" @default.