Abstract :

A Lot Of Research Is Being Done On The Area Where DL And Predicting The Financial Markets Meet. This Is Because The Stock Market Is Very Complex And Is Affected By Many Factors. To Make Good Predictions, You Have To Think About How Volatile The Market Is, Since Stock Prices Quickly Reflect All The Information That Is Available. This Project Suggests A New Way To Predict How Stock Prices Will Move By Using Both Stock And News Data. The Main Goal Is To Create A Strong Predictive Model That Can Understand The Complex Connection Between Stock Prices And Textual News Data. Because Stock Markets Work So Well, The Goal Is To Make Predictions More Accurate By Using A Variety Of Data Sources And Taking Market Instability Into Account. Our Method Involves Making A Hybrid Information Mixing Module That Uses Two Mapblocks To Make The Interaction Between Price And Text Data Features Work Well. This Module Pulls Out The Multimodal Relationships Between Price Series Data Over Time And Semantic Features From News Data. Then, A Multilayer Perceptron-based Model Is Used To Guess How The Stock Price Will Move. Using The Proposed Method In Experiments That Took Place In A Very Volatile Stock Market Showed Promising Outcomes. We Saw Big Improvements In The Accuracy Of Our Predictions When We Combined Price And Text Data Features And Used The Hybrid Information Mixing Module. This ALSO Used ARIMA And A Mix Of LSTM, GRU, And Bidirectional Models, Which Made It Much Better At Making Predictions. With Real-time Data, This Variety Of Approaches Led To A Stronger And More Accurate Final Prediction. “Index Terms - Stock Movement Prediction, Time-series Forecasting, Bidirectional Encoder Representations From Transformer (BERT), Gated Recurrent Units (GRU), Multilayer Perceptron (MLP)”.

Published:

11-9-2025

Issue:

Vol. 25 No. 9 (2025)

Page Nos:

36-47

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite