StockGuru: Smart Way to Predict Stock Price Using Machine Learning

Abstract

Stock price prediction is a trending concepts  in today’s world. Proposed work use Twitter data to predict  public mood and use the predicted mood to predict the stock  market movements. The ceaseless use of social media in the  contemporary era has reached unprecedented levels, which  has led to the belief that the expressed public sentiment could be correlated with the  behaviour of stock prices. Here we develop a system which  collects past tweets, processes them further, and examines  the electiveness of various machine learning techniques such  as Naive Bayes classification and XgBoost algorithm, for  providing a positive, negative or neutral sentiment on the  tweet corpus. Subsequently, work employ an  equivalent machine learning algorithms to analyse how tweets correlate with stock market price behaviour. Finally,  examine our prediction’s error by comparing our algorithm’s  outcome with next day’s actual close price. Here proposed  work takes data from Twitter and also to improve the  accuracy proposed work also takes stock data from  newspapers and yahoo finance also. The final results seem to  be promising as we found correlation between sentiment of  tweets and stock prices.