Analysis of News Sentiment for Stock Price Prediction Using VADER Sentiment

Stock market trading is often thought to be a lucrative subject and in it, is predominantly occupied by the discussion of stock price prediction. Within it comes a multitude of theories pertaining to stock movement, while they seem irreconcilable an integrated model can be created to improve the accuracy of stock prediction as well as form a synthesis of said multitude. The paper focuses on finding the correlation that lies between the historical stock price of Bank Jago alongside economic indicators and news sentiments through implementing VADER sentiment analysis and regression models. The analysis and modelling data were collected in real-time through Yahoo Finance API and GDELT for the stock news. The analysis results indicate that Pearson and Spearman correlation coefficients between stock prices and news sentiments have low values or relationships due the fact that these methods only describe general statistical relationship on average. However, there are reasonably improvements in correlation coefficients when using rolling techniques and economic indicators. In other hand, the stock traders and investors prefer to obtain real-time predictions or indicators of low or high stock prices and the dynamical stock price directions. The latter is denoted by reasonably good dynamical alignments between stock prices and news sentiments through visual inspection of the chart. This research is highly recommended to continue for the next steps by integrating sentiment analysis and machine learning models for predicting stock prices.