A comparative analysis on question classification task based on deep learning approaches

• DOI: 10.7717/peerj-cs.570
• Publication Date: 2021-08-03T09:06:49Z
• Authors: Muhammad Zulqarnain; Ahmed Khalaf Zage...; Rozaida Ghazali; Muhammad Ghulam G...; Wareesa Sharif; Noor Aida Husaini
• Keywords: Word2vec; Word embedding; Sentiment Analysis

Question classification is one of the essential tasks for automatic question answering implementation in natural language processing (NLP). Recently, there have been several text-mining issues such as text classification, document categorization, web mining, sentiment analysis, and spam filtering that successfully achieved by deep learning approaches. In this study, we illustrated investigated our work on certain approaches an extremely inflected Turkish language. trained tested architectures questions dataset Turkish. addition to this, used three main (Gated Recurrent Unit (GRU), Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN)) also applied two different combinations CNN-GRU CNN-LSTM architectures. Furthermore, Word2vec technique with both skip-gram CBOW methods word embedding various vector sizes a large corpus composed user questions. By comparing conducted experiment based test 10-cross fold validation accuracy. Experiment results were obtained illustrate effectiveness techniques considerable impact accuracy rate using We attained 93.7% these dataset.