Ms. Anchal Jain, Neha Goyal, Shruti Mittal
The stock market plays a pivotal role in driving economic growth, where the central objective of investors is to maximize returns while minimizing associated risks. In recent years, Artificial Neural Networks (ANNs) have gained considerable attention as a soft computing technique for stock market forecasting. This study presents a systematic review of ANN-based forecasting models, focusing on their methodologies, datasets, and performance metrics. A total of 289 peer-reviewed journal articles published between 2000 and 2022 were initially retrieved from the Scopus database, out of which 37 were selected based on clearly defined research questions and inclusion criteria. Special emphasis is placed on Long Short-Term Memory (LSTM) networks, which have emerged as one of the most effective architectures for capturing temporal dependencies in stock price movements. The study also provides a bibliometric analysis to identify influential publications, authors, and emerging themes in the field. Despite notable advancements, significant challenges remain in accurately modelling the stock market due to its inherent complexity and volatility. Our findings underscore that stock market forecasting requires a nuanced approach where some variables and modelling strategies deserve more weight than others. This review offers a comprehensive classification of existing studies and highlights future research directions to improve prediction accuracy and model interpretability.