Predicting the Stock Price of Frontier Markets Using LSTM model

Research Supervisor
Journal Title
Journal Title
Journal ISSN
Volume Title
Traditionally most machine learning (ML) models utilize as input highlights a few perceptions (tests / cases) but there's no time measurement within the information. Time-series estimating models are the models that are competent to foresee future values based on previously observed values. Time-series forecasting is widely used for non-stationary data. Non-stationary data are called the data whose statistical properties e.g. the mean and standard deviation are not constant over time but instead, these metrics vary over time. These non stationary input data (used as input to these models) are usually called time-series. A few cases of time-series incorporate the temperature values over time, stock price over time, and cost of a house over time etc. So, the input could be a flag (time-series) that's characterized by perceptions taken successively in time. In this work, using LSTM, we have come up with a comparative analytical approach and numerical procedure to discover the cost of call choice and put alternative and considered these two costs as buying cost and offering cost of stocks of frontier markets so that able to forecast the stock cost (closing price). Changes have been made to the model to find the parameters of previous 60 days closing price to find out the accuracy of the model. To verify the result obtained using modified LSTM we have used deep learning approach using Google colab where we have adopted different algorithms like the decision tree, random forest method and long short-term memory. It has been observed that after trying all the 3 algorithmic approach LSTM has given the best accuracy. Also we’ve found that many companies price fluctuates quite often and with a high margin which makes it quite hard to predict the future price of those companies.
Department Name
Electrical and Computer Engineering
North South University
Printed Thesis