Twitter review analysis and sarcasm detection | ||||
النشرة المعلوماتية في الحاسبات والمعلومات | ||||
Article 3, Volume 1, Issue 3 - Serial Number 20190103, October 2019, Page 12-19 PDF (523.18 K) | ||||
Document Type: المقالة الأصلية | ||||
DOI: 10.21608/fcihib.2019.107520 | ||||
View on SCiNiTO | ||||
Authors | ||||
Mo'men Mohamed; Ahmed Magdy; Shymaa Hussein; Ahmed Samir; Hala Masoud; Hana Morsy; Soha Ahmed Ehssan Aly Mohamed | ||||
Lecturer, Helwan University | ||||
Abstract | ||||
Due to the increase in the number of users on the web, and the increase of the number of reviews that the user's giveaway, it becomes essential to understand and analyze this data. This paper provides a review analysis model for getting feedback from users about specific products found in tweets. This model predicts the polarity of tweet reviews. The main idea of this system is to give report with percent of positive and negative opinions about a specific product. Machine learning (ML) and Natural Language Processing (NLP) approaches are used to get a preliminary determination of the polarity of a tweet by analyzing public ones published on Twitter. In addition, this proposed model uses two techniques: topic modeling and word weight as a feature engineering and three ML algorithms: support vector machine, convolutional neural network (CNN) and naïve bays. The accuracy results of the three algorithms are compared to decide which one is better when using the same data-sets As a conclusion our model aims to provide a whole feedback picture about any product on the social network, but we will use here twitter because it is one of the most popular SN. | ||||
References | ||||
| ||||
Statistics Article View: 278 PDF Download: 277 |
||||