@article{MAKHILLJEAS2020151019299,
    title = {A Sentiment Analysis Approach Based on User Ranking using Type-2 Fuzzy Logic Suitable
for Online Social Networks},
    journal = {Journal of Engineering and Applied Sciences},
    volume = {15},
    number = {10},
    pages = {2315-2326},
    year = {2020},
    issn = {1816-949x},
    doi = {jeasci.2020.2315.2326},
    url = {https://makhillpublications.co/view-article.php?issn=1816-949x&doi=jeasci.2020.2315.2326},
    author = {Magda},
    keywords = {Social Network Analysis (SNA),Sentiment Analysis (SA),Twitter,influential users,Fuzzy Logic (FL),Artificial Neural Networks (ANN)},
    abstract = {The increasable usage of social media in
expressing opinions has raised the importance of Social
Network Analysis (SNA). Business owners utilize SNA
to detect influence users who can motivate others to buy
their products through growing positive feedback. This
emphasizes the need to consider people’s perspectives in
the process of Sentiment Analysis (SA). Considering
perspectivism while computing text polarity can help the
machine to reflect the human perceived sentiment within
text content. Moreover, text vagueness still distresses the
accuracy of SA. In this study, a fuzzy-based SA approach
for Twitter is proposed that handles perspectivism through
integrating SNA with the sentiment process. SA is done
using Text Blob and Fuzzy logic while SNA is done using
UCINET tool and Artificial Neural Networks (ANN) to
rank users. This research aims to avoid misleading
sentiments, improve sentiment classification accuracy,
and deal with social behaviors. After all, a more real
sentiment is produced that reflects what readers have
perceived. The fuzzy classification technique was adopted
to deal with the vagueness of language and for finegrained
classification of Tweets into seven classes instead
of the binary classification. A comparative analysis
between Type-1 and Type-2 fuzzy logic is conducted to
choose the technique with better performance. The
proposed model is practiced on data collected from
Twitter. Results show significance in the use of Type-2
fuzzy logic in terms of model accuracy with the ability to
handle perspectivism.}
    }