TY - GEN

T1 - Named Entity Extraction from Semi-structured Data Using Machine Learning Algorithms

AU - Mansurova, Madina

AU - Barakhnin, Vladimir

AU - Khibatkhanuly, Yerzhan

AU - Pastushkov, Ilya

PY - 2019/1/1

Y1 - 2019/1/1

N2 - The modern society have been witnessed that intensive development of Internet technologies had followed to information explosion during last decades. This explosion had been expressing by an exponential growth of data volume among the low-quality information. This paper is designed to provide detailed information about some intellectual tools which are support decision taking by automatic knowledge extraction. In the first part of paper, we considered a preprocessing contains morphological analysis of texts. Then we had considered the model of text documents in the form of a hypergraph and implementation of the random walk method to extract semantically close word’s pairs, in other words, pairs that often appears together. Result of calculations is matrix with word affinity coefficients corresponding to each other component of vocabulary vector. In the second part we describe training of neural network for linguistic constructions extraction. These ones include possible values of text named entities descriptors. The neural network enables to retrieve information on one preselected descriptor, for example, location, in the form of the final result of the name of geographical objects. In a general case, the neural network can retrieve information on several descriptors simultaneously.

AB - The modern society have been witnessed that intensive development of Internet technologies had followed to information explosion during last decades. This explosion had been expressing by an exponential growth of data volume among the low-quality information. This paper is designed to provide detailed information about some intellectual tools which are support decision taking by automatic knowledge extraction. In the first part of paper, we considered a preprocessing contains morphological analysis of texts. Then we had considered the model of text documents in the form of a hypergraph and implementation of the random walk method to extract semantically close word’s pairs, in other words, pairs that often appears together. Result of calculations is matrix with word affinity coefficients corresponding to each other component of vocabulary vector. In the second part we describe training of neural network for linguistic constructions extraction. These ones include possible values of text named entities descriptors. The neural network enables to retrieve information on one preselected descriptor, for example, location, in the form of the final result of the name of geographical objects. In a general case, the neural network can retrieve information on several descriptors simultaneously.

KW - Entity extraction

KW - Machine learning algorithms

KW - Neural networks

KW - Random walk method

KW - Semi-structured data

UR - http://www.scopus.com/inward/record.url?scp=85072857642&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-28374-2_6

DO - 10.1007/978-3-030-28374-2_6

M3 - Conference contribution

AN - SCOPUS:85072857642

SN - 9783030283735

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 58

EP - 69

BT - Computational Collective Intelligence - 11th International Conference, ICCCI 2019, Proceedings

A2 - Nguyen, Ngoc Thanh

A2 - Chbeir, Richard

A2 - Exposito, Ernesto

A2 - Aniorté, Philippe

A2 - Trawinski, Bogdan

A2 - Nguyen, Ngoc Thanh

PB - Springer-Verlag GmbH and Co. KG

T2 - 11th International Conference on Computational Collective Intelligence, ICCCI 2019

Y2 - 4 September 2019 through 6 September 2019

ER -