Twitter-based Sentiment Analysis of Cryptocurrencies

Abstract

Digital currency has become increasingly popular in recent years. Cryptocurrencies were the first digital currency to be widely used as an asset by most people. The main difference between cryptocurrencies and traditional assets is that they have some similarities, yet various differences that make it difficult for many people to make a decision on whether they want to dive into the world of cryptocurrencies. This research has been performed by scanning Twint so that it used to determine whether or not the reader wants to enter the cryptocurrency market. Using Twint for scraping, this research has been conducted in order to help people who are interested in cryptocurrencies obtain information about sentiment towards cryptocurrencies. This is accomplished through lexicon opinion that determine the direction and the sentiment of the text. This grouping results tested into LSTM model and obtained 95% accurate.