Twitter Sentiment Analysis and Expert Ratings of Initial Coin Offering Fundraising: Evidence from Australia and Singapore Markets

Abstract

Sentiment analysis of Twitter data is quite valuable for determining the market opinion. Twitter sentiment analysis is more challenging than generic sentiment analysis owing to slang and misspellings. The techniques utilized for evaluating the sentiment of tweets that have the greatest importance for the success of an Initial Coin Offering (ICO) are machine learning approaches. In this study, we examined market sentiment and used Expert Ratings to predict the success of ICOs in the Australian and Singapore markets. Based on 68,281 tweets from 57 ICOs across four industries: business services, cryptocurrency, entertainment, and platform. Several classification methods were investigated, including Support Vector Machines (SVMs), Logistic Regression (LR), Random Forest (RF), and Naïve Bayes (NB). The outcomes indicated that sentiment analysis of tweets and expert ratings may be used to forecast the success of an initial coin offering. The results indicate that the suggested model is capable of accurately assessing the tweets of the ICO Successful with a maximum accuracy of about 94.7 % when implementing the Support Vector Machines (SVMs) classifier.