A comparative study on effect of news sentiment on stock price prediction with deep learning architecture

• DOI: 10.1371/journal.pone.0284695
• Publication Date: 2023-04-25T18:23:31Z
• Authors: Keshab Raj Dahal; Nawa Raj Pokhrel; Santosh Gaire; Sharad Mahatara; Rajendra P. Joshi; Ankrit Gupta; Huta R. Banjade; Jeorge Joshi
• Keywords: Stock (firearms)

The accelerated progress in artificial intelligence encourages sophisticated deep learning methods predicting stock prices. In the meantime, easy accessibility of market palm one's hand has made its behavior more fuzzy, volatile, and complex than ever. world is looking at an accurate reliable model that uses text numerical data which better represents market's highly volatile non-linear a broader spectrum. A research gap exists accurately target stock's closing price utilizing combined data. This study long short-term memory (LSTM) gated recurrent unit (GRU) to predict using features alone incorporating financial news conjunction with features. comparative carried out under identical conditions dispassionately evaluates importance prediction. Our experiment concludes produces prediction accuracy fundamental alone. performances architecture are compared standard assessment metrics -Root Mean Square Error (RMSE), Absolute Percentage (MAPE), Correlation Coefficient (R). Furthermore, statistical tests conducted further verify models' robustness reliability.