Premium
Semi‐supervised learning induced abnormal emotional tendency analysis in British culture
Author(s) -
Ai Qing,
Zhou Ding
Publication year - 2020
Publication title -
internet technology letters
Language(s) - English
Resource type - Journals
ISSN - 2476-1508
DOI - 10.1002/itl2.239
Subject(s) - overtraining , sentiment analysis , computer science , focus (optics) , artificial intelligence , natural language processing , psychology , speech recognition , cognitive psychology , medicine , physics , optics , athletes , physical therapy
The abnormal emotional tendency analysis refers to the automatically detecting the minor negative sentiment from various emotions which are obtained from text, voice, image, or video. In this paper, we focus on detecting abnormal emotional tendency in English sentences on Twitter. In order to solve the issue that the text in Twitter is fragment and the associated emotions are imbalance, we extract proper features of Twitter to represent abnormal emotional tendency and input the extracted features into an improved self‐supervised reserved self‐training. Then, we optimize the termination condition by using a threshold method which is based on sentiment tendency analysis. The optimized termination condition cannot only preserve the merit that reserved self‐training can solve the imbalance emotions in Twitter corpus, but also can resist the overtraining. The experimental results show that the improved semi‐supervised reserved self‐training is effect.