Word2vec convolutional neural networks for classification of news articles and tweets

• DOI: 10.1371/journal.pone.0220976
• Publication Date: 2019-08-22T17:44:30Z
• Authors: Beakcheol Jang; Inhwan Kim; Jong Wook Kim
• Keywords: Word2vec; Word embedding; n-gram; Bag-of-words model; chEMBL; Sentiment Analysis

Big web data from sources including online news and Twitter are good resources for investigating deep learning. However, collected articles tweets almost certainly contain unnecessary learning, this disturbs accurate This paper explores the performance of word2vec Convolutional Neural Networks (CNNs) to classify into related unrelated ones. Using two word embedding algorithms word2vec, Continuous Bag-of-Word (CBOW) Skip-gram, we constructed CNN with CBOW model Skip-gram model. We measured classification accuracy CBOW, without models real tweets. The experimental results indicated that significantly improved was higher more stable when compared exhibited better on articles, Specifically, effective because typically uniform