Special issue on "The Eighth International Symposium on Information and Communication Technology—SoICT 2017".

First, TD2V, a generic pre-trained document embedding for English documents from more than one million tweets in Twitter is trained, then, domain adaptation process with adversarial training is applied to adapt TD2V to different domains. Experiments on datasets from different online sources show that by using TD2V only, the proposed method can classify documents with better accuracy than existing methods. B Introductory text: b Since 2010, the Symposium on Information and Communication Technology - SoICT has been organized annually. [Extracted from the article]