Analysis of Cryptocurrency PublicSentiment Shifts

Abstract

The popularity and emergence of digital currency can be attributed to social media as it has a large user base for online discussion. The crypto market is highly dependent on socially constructed opinions as investors rely on online sources to acquire related information. A headline is considered as a summary of an article in a single sentence as it is the first that can gain a reader’s attention. The objective of this paper is to analyze the trend of cryptocurrency headlines in terms of positive and negative sentiment. The dataset is obtained from the Kaggle webpage containing news headlines from five online platforms. The paper will be using a lexicon-based sentiment approach to identify the binary sentiment. Support Vector Machine algorithm will be used to evaluate and optimize the sentiment results. This paper contributes to the gap in the literature by providing an empirical analysis of overview changes of the cryptocurrencies with multiple news platforms and longer periods. The findings of the paper resulting in negative sentiment for the trend of headlines, but it showed a different view in terms of positive polarity. Furthermore, the TD-IDF model outperformed the sentiment model in SVM modeling. Keywords: Cryptocurrency, Sentiment analysis, Lexicon sentiment, Natural Language Processing (NLP), Support Vector Machine (SVM).