Detail publikace

Linguistically independent sentiment analysis using convolutional-recurrent neural networks model

Originální název

Linguistically independent sentiment analysis using convolutional-recurrent neural networks model

Anglický název

Linguistically independent sentiment analysis using convolutional-recurrent neural networks model

Jazyk

en

Originální abstrakt

Text classification is a process which analyses text and assigns one or more classes to it based on its content. This paper introduces a linguistically independent text classifier based on convolutional–recurrent neural networks. The classifier works at character level instead of some higher structures such as words, sentences, etc. To evaluate the accuracy of the proposed methodology, the Yelp data set and other multilingual data set obtained from film review databases containing Czech, German and Spanish languages were used. The resulting accuracy on the Yelp data set is 93,64 %. We also proved that the proposed model can work for various languages.

Anglický abstrakt

Text classification is a process which analyses text and assigns one or more classes to it based on its content. This paper introduces a linguistically independent text classifier based on convolutional–recurrent neural networks. The classifier works at character level instead of some higher structures such as words, sentences, etc. To evaluate the accuracy of the proposed methodology, the Yelp data set and other multilingual data set obtained from film review databases containing Czech, German and Spanish languages were used. The resulting accuracy on the Yelp data set is 93,64 %. We also proved that the proposed model can work for various languages.

BibTex


@inproceedings{BUT157766,
  author="Vojtěch {Myška} and Radim {Burget} and Lukáš {Povoda} and Malay Kishore {Dutta}",
  title="Linguistically independent sentiment analysis using convolutional-recurrent neural networks model",
  annote="Text classification is a process which analyses text and assigns one or more classes to it based on its content. This paper introduces a linguistically independent text classifier based on convolutional–recurrent neural networks. The classifier works at character level instead of some higher structures such as words, sentences, etc. To evaluate the accuracy of the proposed methodology, the Yelp data set and other multilingual data set obtained from film review databases containing Czech, German and Spanish languages were used. The resulting accuracy on the Yelp data set is 93,64 %. We also proved that the proposed model can work for various languages.",
  address="IEEE",
  booktitle="2019 42nd International Conference on Telecommunications and Signal Processing (TSP)",
  chapter="157766",
  doi="10.1109/TSP.2019.8768887",
  howpublished="online",
  institution="IEEE",
  year="2019",
  month="july",
  pages="212--215",
  publisher="IEEE",
  type="conference paper"
}