Abstract
The increase in the volume of user-generated content on Twitter has resulted in tweet sentiment analysis becoming an essential tool for the extraction of information about Twitter users' emotional state. Consequently, there has been a rapid growth of tweet sentiment analysis in the area of natural language processing. Tweet sentiment analysis is increasingly applied in many areas, such as decision support systems and recommendation systems. Therefore, improving the accuracy of tweet sentiment analysis has become practical and an area of interest for many researchers. Many approaches have tried to improve the performance of tweet sentiment analysis methods by using the feature ensemble method. However, most of the previous methods attempted to model the syntactic information of words without considering the sentiment context of these words. Besides, the positioning of words and the impact of phrases containing fuzzy sentiment have not been mentioned in many studies. This study proposed a new approach based on a feature ensemble model related to tweets containing fuzzy sentiment by taking into account elements such as lexical, word-type, semantic, position, and sentiment polarity of words. The proposed method has been experimented on with real data, and the result proves effective in improving the performance of tweet sentiment analysis in terms of the $F_{1}$ score.
Highlights
With the growth of social networks, an increasing number of people want to find, share, and exchange information with each other without any regard to the geographical distance
RESEARCH QUESTION To improve the accuracy of analyzing sentiment in tweets containing fuzzy sentiment of previous method, the main question for the research is as follows: How can we improve the performance of analyzing the sentiment of tweets containing fuzzy sentiment based on the feature ensemble model? This question is partitioned into the two following sub-questions: The first question: How can a feature ensemble model based on a set of features extracted from tweets be built?
Our proposed method consists of three main steps: 1) a set of features related to tweets containing fuzzy sentiment are extracted; 2) a feature ensemble model to create tweet embeddings is proposed by combining feature vectors extracted in the first step; 3) a CNN model is used to classify the sentiment of tweets into five sets such as negative tweets set, neutral tweets set, positive tweets set, strong positive tweets set, and strong negative tweets set
Summary
With the growth of social networks, an increasing number of people want to find, share, and exchange information with each other without any regard to the geographical distance. The Word2Vec and GloVe models need massive data for training and creating a suitable vector for each word [1], [13] These methods may not be the best conform for small and informal data such as tweets. Many approaches have been tested to improve the accuracy of TSA methods with relatively good results by using the feature ensemble method Most of these methods attempted to model the syntactic information of words while ignoring the sentiment context. There are some researchers who have been tried to build a feature ensemble model, but they have not fully considered the features, such as the lexical, word-type, semantic, position of words They have not yet mentioned the impact of the fuzzy sentiment phrases. The conclusions and future work are discussed in the last section
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
