Modelling Stock Prices Prediction with Long Short-Term Memory (LSTM): A Black Box Approach

A major requirement for small, medium or large investment firms is to be able to invest in the right stock and attain profitable growth. Giant firms use machine learning intelligence to maintain their competitive edge over the others by trying to predict the next profitable stock precisely. Stock price forecasting is a crucial part of these businesses, and this brings up the research questions such as, what will be the next price of the stock? Does the company need to short the stock? Is long-term holding the best approach at the moment? How to cut losses? Among its many benefits, predictive forecast is a key enabler for a better stock investment model. This research paper introspects on various key aspects of the Long Short-Term Memory (LSTM), a deep learning algorithm which focuses on Recurrent Neural Networks (RNNs) and is a black box approach that is apt for predicting stock prices. The paper evaluates the performance of the proposed model and aims to explain the complexity of the system and the need to move towards grey box approach for better comprehension.