STOCK PRICE PREDICTION FROM NEWS HEADLINES USING MACHINE LEARNING MODELS

Authors

  • Ahmed J. Obaid Asst. Professor, Faculty of Computer Science and Mathematics, University of Kufa, Najaf, Iraq

Keywords:

BERT, RNN, Sentiment analysis, Tokenization.

Abstract

It's a tremendously intriguing and exciting issue to forecast and speculate on stock market values, especially worldwide company values. This article uses economic news received from businesses to discuss changes in stock prices and projections of stock values. Pay attention to business news headlines and assess headline sentiment using a number of tactics. The Neural Network reconstructs sentiment outcomes with changes in equities over the same period by using BERT as a benchmark and comparing the findings with three other tools. Compared to the other two tools, BERT and RNN are substantially more accurate since they can recognise emotional values without the neutral component. Establish when changes in stock values occur by contrasting these findings with stock value fluctuations over the same time period using sentiment analysis of economic news articles. The impact of sentimental value on changes in sentimental stock market value was also shown to vary significantly amongst the various models.

Downloads

Published

2024-09-24

Issue

Section

Articles