Análise do impacto de notícias na previsão do índice Bovespa

Abstract

Currently, the principal source of information for Brazilians is social networks. In a society where citizens have easy access to information. The spread of news occurs rapidly. Since, through social networks, information can effortlessly propagate. This work proposes to conduct the IBOV forecast, considering the effects of data transmitted on the internet to contribute to understanding its outcomes. And as a result, minimize damage. Because it is possible to relative news and tweet content with oscillations in the financial market. The predictive models developed utilizing techniques derived from the Box-Jenkins class. And enriched with the sentiment values of news and tweets. The models that achieved the best results considered the sentiments of tweets. From 2018 to 2022, the daily forecast model that considered the entire historical series of the closing price of the IBOV, performed better considering only the sentiments of tweets. The prototype is capable of forecasting with an average difference between the predicted value and the actual value of approximately 0.0066 of MAPE.