Enhancing Stock Price Predictions with Joint Network Models: A Biotechnological Approach to Financial Analytics

Abstract

The volatility and unpredictability of the stock market make it a significant area of interest for investors and investment firms. Traditional and contemporary methods to predict stock trends and prices have typically been segmented into statistical techniques, such as logistic regression and ARCH models, and artificial intelligence (AI) methodologies, including multilayer perceptrons, convolutional neural networks, and Bayesian networks, among others. While these methods have been somewhat effective, they often restrict predictions to single value outputs, limiting their practical utility for dynamic stock market analysis. Addressing this limitation, this paper introduces an innovative joint network model employing a deep recurrent neural network architecture based on Long Short-Term Memory (LSTM) networks that accommodates multiple inputs and outputs. This model is designed to predict several interrelated stock price indicators—opening, low, and high prices—simultaneously. This capability enhances the model's utility by providing a comprehensive view of stock price movements, which is crucial for making informed investment decisions. Comparative analyses with standalone LSTM network models and traditional deep recurrent neural networks demonstrate the superior performance of our proposed joint network model. Experimental results reveal that our model achieves a prediction accuracy of over 95%, outperforming existing models in simultaneous multiple value predictions. This advancement not only underscores the model's efficacy but also highlights its potential applications in biotechnological data processing and financial analytics, where precision and multidimensional analysis are paramount.