Comparative analysis of methods for solving the problem of sentiment text analysis.
Abstract
This article discusses a supervised learning approach for sentiment analysis of text data, using NLP. The research was conducted, using four methods on the same dataset and their efficiency was compared by using the following characteristics: training time, testing time and accuracy of classification. A 3D CNN model using BERT Tokenizer for text preprocessing was selected as the best method of this comparative study, because of its text preprocessing algorithm.
References
Bing Liu. (2012) Sentiment Analysis and Opinion Mining, Morgan & Claypool Publishers, May 2012.
Kobec N.M., Kovalyuk T.V. (2018) Zadachi sentiment-analizu yak instrument monitoringu mirkuvan koristuvachiv socialnih merezh.Materiali naukovoyi konferenciyi studentiv, magistrantiv ta aspirantiv 23 – 24 kvitnya 2018 roku «Informatika ta obchislyuvalna tehnika» – IOT-2018, Kiyiv. S.151-153.
Maznichenko N. I. (2017) Avtomatichnij analiz tekstiv elektronnih dokumentiv. Suchasni tendenciyi rozvitku ukrayinskoyi nauki : materiali Vseukr. nauk. konf., 6–7 trav. 2017 r.. Pereyaslav-Hmelnickij. Vip. 2. S. 136–139.
Shingalov D. V. (2017) Metodi avtomatichnogo analizu tonalnosti kontentu u socialnih merezhah dlya viyavlennya informacijno-psihologichnih vpliviv / D. V. Shingalov, Ye. V. Meleshko, R. M. Minajlenko, V. A. Reznichenko // Zbirnik naukovih prac Kirovogradskogo nacionalnogo tehnichnogo universitetu. Tehnika v silskogospodarskomu virobnictvi, galuzeve mashinobuduvannya, avtomatizaciya. Vip. 30. S. 196-202. Rezhim dostupu: http://nbuv.gov.ua/UJRN/znpkntu_2017_30_29.
Shahovska N. B. (2017) Shkalyuvannya emocijno zabarvlenih sliv dlya vikoristannya u metodah klasifikaciyi tonalnosti / N. B. Shahovska, H. Yu. Girak // Visnik Nacionalnogo universitetu "Lvivska politehnika". Informacijni sistemi ta merezhi. № 872. S. 195-203. Rezhim dostupu: http://nbuv.gov.ua/UJRN/VNULPICM_2017_872_23.
Large Movie Review Dataset [Elektronnij resurs]. - URL: https://www.kaggle.com/lakshmi25npathi/imdb-dataset-of-50k-movie-reviews
Stuart J. (2003) Russell and Peter Norvig. Artificial Intelligence: A Modern Approach, Prentice Hall.
Understanding RNN and LSTM [Elektronnij resurs]. – URL: https://towardsdatascience.com/understanding-rnn-and-lstm-f7cdf6dfc14e
Primenenie svyortochnyh nejronnyh setej dlya zadach NLP [Elektronnij resurs] – URL: https://habr.com/ru/company/ods/blog/353060/
TensorFlow [Elektronnij resurs] – URL: https://en.wikipedia.org/wiki/TensorFlow
Keras [Elektronnij resurs] – URL: https://en.wikipedia.org/wiki/Keras
Understanding of Convolutional Neural Network (CNN) — Deep Learning [Elektronnij resurs] – URL: https://medium.com/@RaghavPrabhu/understanding-of-convolutional-neural-network-cnn-deep-learning-99760835f148
Sentiment Analysis: Concept, Analysis and Applications [Elektronnij resurs] – URL: https://towardsdatascience.com/sentiment-analysis-concept-analysis-and-applications-6c94d6f58c17
Abstract views: 0 PDF Downloads: 0
