Abstract:
Analyzing sentiments has been widely regarded as a popular technique by many researchers, and Twitter nominated the most user-friendly, and reliable social media supplying the stream of sentiments. Among the trendiest topics of discussion in such social platforms, cryptocurrency, and most notably Bitcoin ranks the highest, both providing curiosity as a technology, and a lucrative asset to trade. This thesis studies the correlation among user sentiments from Twitter and the change in price of Bitcoin, to carve out a scalable model by manipulating the category of sentiments as variables and appropriate quantitative machine learning techniques.
The work finally achieved a stable precision for determining movement in price, with a high of 75% in accuracy in the short run.
