The understanding of financial behavior, especially its application in the stock market, has become increasingly important in recent years due to its significant impact on the global economy. One field that explores the intersection of finance and computer science to create predictive models is known as stock market prediction. This field aims to predict the behavior of various securities in the financial market. Deep Learning, one of the most renowned and utilized techniques, consists of various deep neural network structures that facilitate learning from non-linear models. In this study, we utilized open data from some of Brazil’s largest companies – Petrobras (PETR4), Itausa (ITSA4), and Vale (VALE3) – provided by BovDB, which includes stock quote data for all companies listed on the Brazilian stock market from 2000 to 2020. The data for the years in question were processed using a recurrent neural network to assess the impact of a price correction factor that accounts for the influence of past events not included in the training and validation results of the RNN model. The findings indicate a strong correlation of the model with temporal data and suggest a positive effect on reducing noise and forecast errors during model training.