Abstract:Stock price forecasting is a crucial area of research that demands a thorough comprehension of market dynamics and sophisticated analytical methods. Deep neural networks have recently demonstrated considerable potential for enabling academics to create extremely precise models for foretelling financial patterns. Another quickly developing technology is Natural language processing (NLP), which is increasingly used to evaluate financial data, notably news, and social media sentiment, and forecast the course of the market. An innovative deep learning-based stock market prediction model is created in this research paper. The historic stock market data and news feed as the information source. The acquired raw data (news data) are pre-processed using lowercase text conversion, punctuation removal, stop word removal, tokenization of sentences, normalization, and a bag of words technique. Then, from the pre-processed news data, the features such as Parts of Speech (PoS), Term Frequency-Inverse Document Frequency (TF-IDF), and N-gram-based features are extracted. In addition, the Moving Average Convergence/Divergence oscillator (MACD), and Relative Strength Index (RSI) based features are extracted from the historic stock market data. The extracted features from the historic stock market data and news feeds are fused. The optimal features are then chosen from the fused features using a new hybrid optimization model called Exchange Market Emperor Pigeon Optimization (EMEPO). The proposed EMEPO model is a combination of the standard Exchange Market Algorithm (EMA) and Emperor Pigeon Optimization (EPO). A new three-phase classifier is introduced in this research work, for accurate stock price forecasting. The proposed three-phase classifier includes Convolutional neural networks (CNN), Bidirectional long short-term memories (Bi-LSTM), and Autoencoder. The three-phase classifier is trained using the selected EMEPO-based features. The projected outcome from the three-phase classifier portrays the stock market prices. In all three feature selection events, the suggested method produces positive outcomes. In Case 1 (Open/Close), the Mean Absolute Error (MAE) value is 0.051582, suggesting a minimal average difference between the anticipated and actual values. The accuracy of the model in forecasting percentage errors is shown by the Mean Absolute Percentage Error (MAPE), which is 11.28154%. Similar results are obtained by the model in Case 2 (High/Low/Close), with an MAE of 0.056408 and a MAPE of 12.0081%. These results show how well the suggested strategy works for anticipating stock prices with accuracy.

BOOST STOCK FORECASTING ACCURACY USING THE MODIFIED FIREFLY ALGORITHM AND MULTICHANNEL CONVOLUTIONAL NEURAL NETWORK

Stock Price Prediction Using Convolutional Neural Networks on a Multivariate Timeseries

Stock Price Prediction using Multi-Faceted Information based on Deep Recurrent Neural Networks

Predicting Stock Market Trends with a Transfer Learning Model Based on CNN-LSTM Architecture

Enhancing stock market prediction using three-phase classifier and EM-EPO optimization with news feeds and historical data

Predicting Stock Closing Price with Stock Network Public Opinion Based on AdaBoost-AAFSA-Elman Model and CEEMDAN Algorithm

An Improved Index Price/Movement Prediction by using Ensemble CNN and DNN Deep Learning Technique

A novel multi-source information-fusion predictive framework based on deep neural networks for accuracy enhancement in stock market prediction

Enhanced stock market forecasting using dandelion optimization-driven 3D-CNN-GRU classification

Convolutional neural network for stock trading using technical indicators

Stock Price Prediction Using CNN and LSTM-Based Deep Learning Models

Prediction of stock market using grey wolf optimization with hybrid convolutional neural network and bi-directional long-short term memory model

Modeling of Hyperparameter Tuned Hybrid CNN and LSTM for Prediction Model

LT-SMF: long term stock market price trend prediction using optimal hybrid machine learning technique

A Multi Parameter Forecasting for Stock Time Series Data Using LSTM and Deep Learning Model

Research on the application of deep learning techniques in stock market prediction and investment decision-making in financial management

Forecasting price in a new hybrid neural network model with machine learning

Forecasting significant stock price changes using neural networks

Genetic algorithm-optimized multi-channel convolutional neural network for stock market prediction

Stock Price Prediction Using CNN-BiLSTM-Attention Model

Stock Price Prediction Using Deep Learning Models