Analysis Of Stock Prediction Parameters and Their Impact on Effective Selection Of Stock

Abstract

Stock value prediction is a multi-disciplinary field which requires efficient knowledge about the stock’s historic values, its news feeds, twitter sentiments, impact of global stock market(s) on the stock, etc. In order to effectively analyse a stock’s trend for inter-day, intra-day or long term, analysists have to evaluate these values on a continuous basis. Along with these values analysists also have to analyse non-stock data like recent news about the company, management changes in the company, tweets related to the company, global news & global stock market trends which affect the company in any way possible. Each of these data sources have a different effect on the stock’s value change, and it is recommended for a good stock prediction system to analyse the effects of these values before real-time deployment. Neglecting even a single parameter before deployment of the stock prediction system might result into a multitude of prediction errors. For instance, if twitter feeds for a stock are not considered during prediction of a nicely performing stock, and suddenly some news about a product fail comes online, then the stock prices might plummet, and the system will not be able to track it. In order to reduce the effect of these outlier events on predicted value of the stock, this paper analyses different parameters that affect stock prediction, and suggest the impact of these parameters on stock performance. Researchers can use this information in order to improve the accuracy of their deployed systems, and make these systems future proof.