Stock Market Forecasting

Abstract

Forecasting stock market prices has been a topic of interest lately, especially, among the researchers and the analysts. But forecasting/predicting stock prices does not come in handy as it is a pretty complex task. The stock market is a highly volatile concept as it is affected by a number of factors viz. Previous performance of the stocks, political factors, investor’s sentiment, change in leadership, etc. It has been observed that historical data and previous performances of the stocks have been inefficient in forecasting the accurate nature of stock.

Existing studies focusing in or around stock market forecasting, focuses on only one aspect/parameter i.e. Historical Data/Previous Performance. There are many algorithms but, even though the accuracy rate is high for some algorithms, one cannot be completely certain that he/she can invest based on only one behaviour of the stock market. Using the sentiment analysis on the tweets collected using the Twitter API and the closing value of various stocks, one can build a handy system that can forecast the stock price movement of various companies various days prior. This paper is based on the above mentioned methodologies and generates a decent accuracy rate which can be further worked upon for more improved results.

Keywords: forecasting stock prices, historical data, sentiment analysis, twitter API