SENTIMENTAL ANALYSIS OF TWITTER: TRENDS & INSIGHTS

Abstract

This paper deals with the sentiment analysis on Twitter. It tackles the problem on community identification using sentiment analysis and neural networks. We sift through and prepare a vast collection of tweets to use, then we apply the text on a sentiment analysis to try and determine the feeling provoked in the tweet. We then proceed to construct a neural network that will seek to look for some patterns in the sentiment data and cluster users into communities according to how they express themselves emotionally. Our method reconstructs the communities using the preceding sentiment profiles of the preceding sentiment profiles of the particular users which add up to tree having similar structure. Such results are useful by showing the community model and how the community operates. The nature of the problem has been addressed by showing the effectiveness of the method in identifying communities in consistent manners on Twitter while defeating standard approaches. The findings of this research will advance the knowledge of online social networks, sentiment contagion, and opinion construction.