Before a certain topic becomes a very searched subject on news platform, there are some weak signals that, if correctly recognized and handled, may anticipated the popularity of that topic. One big problem with detecting such weak signals is that their recognition relies to a large extent on human tacit knowledge. Human tacit knowledge is a type of information having as main characteristics the fact that there is not a direct formal definition of it, and there is not a direct label in the text which explicitly marks it. In this paper we report on building an annotated news corpus for detection of weak signals. We also report on experiments using a supervised machine learning technique.
@InProceedings{IRIMIA18.11, author = {Alina Irimia ,Punguta Paul and Radu Gheorghiu}, title = {Tacit Knowledge - Weak Signal Detection}, booktitle = {Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)}, year = {2018}, month = {may}, date = {7-12}, location = {Miyazaki, Japan}, editor = {Octavian Popescu and Carlo Strapparava}, publisher = {European Language Resources Association (ELRA)}, address = {Paris, France}, isbn = {979-10-95546-11-5}, language = {english} }
Summary of the paper