Stock Market Forecasting with Different Input Indicators using Machine Learning and Deep Learning Techniques: A Review

Machine Learning and Deep Learning-based stock market prediction models are in trend. Various methods and algorithms are there to deal with stock market analysis. However, an efficient forecasting model is tough to design. Researchers always try to develop a better and more efficient forecasting model than the previous one. At the initial stage of any research, the dilemma is to identify the research problem and prospective model to solve the selected issue. Also, one has to consider a different aspect of the problem with shortcomings of the existing research. This review article helps the researchers to get an idea about these aspects. This paper reviews research articles that were relevant to stock market forecasting that used machine learning and deep learning techniques. The selection of research articles is based on the input indicator's usage. Review categorization is based on historical data, technical or fundamental indicator utilization; public sentiment identification and analysis; and other influential factor considerations. A detailed review gives precise information about recent trends in the utilization of input indicators, different machine learning and deep learning methods, datasets, statistical tests, and evaluation metrics. Literature emphasized more on the technical indicator and sentiment analysis-based prediction approach. Financial ratios and influential factors are less considered. Based on the shortcoming of the existing work, it is needed to propose a more effective and accurate prediction model. Optimized deep learning models are trending. So, we suggest a metaheuristic-based optimized deep learning framework for future work that may utilize various possible input indicators.