The impact of directionality in predications on text mining

Gondy Augusta Leroy, Marcelo Fiszman, Thomas C. Rindflesch

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

The number of publications in biomedicine is increasing enormously each year. To help researchers digest the information in these documents, text mining tools are being developed that present co-occurrence relations between concepts. Statistical measures are used to mine interesting subsets of relations. We demonstrate how directionality of these relations affects interestingness. Support and confidence, simple data mining statistics, are used as proxies for interestingness metrics. We first built a test bed of 126,404 directional relations extracted from biomedical abstracts, which we represent as graphs containing a central starting concept and 2 rings of associated relations. We manipulated directionality in four ways and randomly selected 100 starting concepts as a test sample for each graph type. Finally, we calculated the number of relations and their support and confidence. Variation in directionality significantly affected the number of relations as well as the support and confidence of the four graph types.

Original languageEnglish (US)
Title of host publicationProceedings of the Annual Hawaii International Conference on System Sciences
DOIs
StatePublished - 2008
Externally publishedYes
Event41st Annual Hawaii International Conference on System Sciences 2008, HICSS - Big Island, HI, United States
Duration: Jan 7 2008Jan 10 2008

Other

Other41st Annual Hawaii International Conference on System Sciences 2008, HICSS
CountryUnited States
CityBig Island, HI
Period1/7/081/10/08

Fingerprint

Data mining
Statistics

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Leroy, G. A., Fiszman, M., & Rindflesch, T. C. (2008). The impact of directionality in predications on text mining. In Proceedings of the Annual Hawaii International Conference on System Sciences [4438932] https://doi.org/10.1109/HICSS.2008.443

The impact of directionality in predications on text mining. / Leroy, Gondy Augusta; Fiszman, Marcelo; Rindflesch, Thomas C.

Proceedings of the Annual Hawaii International Conference on System Sciences. 2008. 4438932.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Leroy, GA, Fiszman, M & Rindflesch, TC 2008, The impact of directionality in predications on text mining. in Proceedings of the Annual Hawaii International Conference on System Sciences., 4438932, 41st Annual Hawaii International Conference on System Sciences 2008, HICSS, Big Island, HI, United States, 1/7/08. https://doi.org/10.1109/HICSS.2008.443
Leroy GA, Fiszman M, Rindflesch TC. The impact of directionality in predications on text mining. In Proceedings of the Annual Hawaii International Conference on System Sciences. 2008. 4438932 https://doi.org/10.1109/HICSS.2008.443
Leroy, Gondy Augusta ; Fiszman, Marcelo ; Rindflesch, Thomas C. / The impact of directionality in predications on text mining. Proceedings of the Annual Hawaii International Conference on System Sciences. 2008.
@inproceedings{28dd7db46a4e41d0bd4d85a532ca2f55,
title = "The impact of directionality in predications on text mining",
abstract = "The number of publications in biomedicine is increasing enormously each year. To help researchers digest the information in these documents, text mining tools are being developed that present co-occurrence relations between concepts. Statistical measures are used to mine interesting subsets of relations. We demonstrate how directionality of these relations affects interestingness. Support and confidence, simple data mining statistics, are used as proxies for interestingness metrics. We first built a test bed of 126,404 directional relations extracted from biomedical abstracts, which we represent as graphs containing a central starting concept and 2 rings of associated relations. We manipulated directionality in four ways and randomly selected 100 starting concepts as a test sample for each graph type. Finally, we calculated the number of relations and their support and confidence. Variation in directionality significantly affected the number of relations as well as the support and confidence of the four graph types.",
author = "Leroy, {Gondy Augusta} and Marcelo Fiszman and Rindflesch, {Thomas C.}",
year = "2008",
doi = "10.1109/HICSS.2008.443",
language = "English (US)",
isbn = "0769530753",
booktitle = "Proceedings of the Annual Hawaii International Conference on System Sciences",

}

TY - GEN

T1 - The impact of directionality in predications on text mining

AU - Leroy, Gondy Augusta

AU - Fiszman, Marcelo

AU - Rindflesch, Thomas C.

PY - 2008

Y1 - 2008

N2 - The number of publications in biomedicine is increasing enormously each year. To help researchers digest the information in these documents, text mining tools are being developed that present co-occurrence relations between concepts. Statistical measures are used to mine interesting subsets of relations. We demonstrate how directionality of these relations affects interestingness. Support and confidence, simple data mining statistics, are used as proxies for interestingness metrics. We first built a test bed of 126,404 directional relations extracted from biomedical abstracts, which we represent as graphs containing a central starting concept and 2 rings of associated relations. We manipulated directionality in four ways and randomly selected 100 starting concepts as a test sample for each graph type. Finally, we calculated the number of relations and their support and confidence. Variation in directionality significantly affected the number of relations as well as the support and confidence of the four graph types.

AB - The number of publications in biomedicine is increasing enormously each year. To help researchers digest the information in these documents, text mining tools are being developed that present co-occurrence relations between concepts. Statistical measures are used to mine interesting subsets of relations. We demonstrate how directionality of these relations affects interestingness. Support and confidence, simple data mining statistics, are used as proxies for interestingness metrics. We first built a test bed of 126,404 directional relations extracted from biomedical abstracts, which we represent as graphs containing a central starting concept and 2 rings of associated relations. We manipulated directionality in four ways and randomly selected 100 starting concepts as a test sample for each graph type. Finally, we calculated the number of relations and their support and confidence. Variation in directionality significantly affected the number of relations as well as the support and confidence of the four graph types.

UR - http://www.scopus.com/inward/record.url?scp=51549108313&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=51549108313&partnerID=8YFLogxK

U2 - 10.1109/HICSS.2008.443

DO - 10.1109/HICSS.2008.443

M3 - Conference contribution

AN - SCOPUS:51549108313

SN - 0769530753

SN - 9780769530758

BT - Proceedings of the Annual Hawaii International Conference on System Sciences

ER -