TY - GEN
T1 - Legal information retrieval system relevant to R&D projects based on word-embedding of core terms
AU - Jung, Hae Min
AU - Lee, Youna
AU - Kim, Wooju
N1 - Publisher Copyright:
© Copyright 2017 ACM
Copyright:
Copyright 2019 Elsevier B.V., All rights reserved.
PY - 2017/8/17
Y1 - 2017/8/17
N2 - Generally, Research and development projects have relationship with statutes. Sometimes, new technologies developed in R&D projects can't be applied because they are restricted by newly enacted statutes. The situation comes from the fact that researchers don't know well about statutes that might affect their R&D projects. Therefore, we proposed a methodology to find relevant statutes with the R&D plan, based on cosine similarity of document vectors. R&D plan is a document which is written by researchers before they get into R&D, and it contains the main concepts about the project. In our method, a R&D plan is represented as a vector using network centrality. Then cosine similarity between the vector and TF-IDF vectors of statutes is calculated. After the calculation, statutes are provided with their ranks based on similarity values, so researchers can review and check them with priority. Compared to our previous study, this study differs in that we applied new variants of vectors to enhance the performance of our methodology and find a better representation for documents.
AB - Generally, Research and development projects have relationship with statutes. Sometimes, new technologies developed in R&D projects can't be applied because they are restricted by newly enacted statutes. The situation comes from the fact that researchers don't know well about statutes that might affect their R&D projects. Therefore, we proposed a methodology to find relevant statutes with the R&D plan, based on cosine similarity of document vectors. R&D plan is a document which is written by researchers before they get into R&D, and it contains the main concepts about the project. In our method, a R&D plan is represented as a vector using network centrality. Then cosine similarity between the vector and TF-IDF vectors of statutes is calculated. After the calculation, statutes are provided with their ranks based on similarity values, so researchers can review and check them with priority. Compared to our previous study, this study differs in that we applied new variants of vectors to enhance the performance of our methodology and find a better representation for documents.
UR - http://www.scopus.com/inward/record.url?scp=85061280510&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85061280510&partnerID=8YFLogxK
U2 - 10.1145/3154943.3154954
DO - 10.1145/3154943.3154954
M3 - Conference contribution
AN - SCOPUS:85061280510
T3 - ACM International Conference Proceeding Series
BT - Proceedings of the International Conference on Electronic Commerce, ICEC 2017
PB - Association for Computing Machinery
T2 - 2017 International Conference on Electronic Commerce, ICEC 2017
Y2 - 17 August 2017 through 18 August 2017
ER -