To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
Long Short-Term Memory Model for Predicting Stock Prices in Tanzania |
|
Author |
Samuel Joseph, Neema Mduma, and Devotha Nyambo |
| Citation |
Vol. 22 No. 11 pp. 597-601 |
|
Abstract |
Building stock price prediction models helps to provide investors with tools for making better data-based decisions. The models help traders to reduce investment risk and select the most profitable stocks. Different studies have been conducted in various countries to predict stock prices. In particular, machine learning and deep learning have been frequently utilized given their success in various fields such as medical diagnosis and language translation. However, literature is not readily available of efforts to exploit either machine learning or deep learning to predict stock prices in Tanzania. This study was conducted to address this gap. The study selects active companies from the Dar es Salaam Stock Exchange (DSE) and builds a Long Short-Term Memory (LSTM) deep neural network to predict next day’s closing price for each company. Since each company has different number of outstanding shares, the study proposed to use: “Volume/Outstanding Shares”, “Outstanding Bids/Outstanding Shares” and “Outstanding Offers/Outstanding Shares”, among the features, instead of: Volume, Outstanding Bids and Outstanding Offers respectively. Using the proposed approach the study realized an improvement in Root Mean Square Error (RMSE) of 54.62%. Therefore the study contributes to available literature by recommending to take into account the companies’ different number of outstanding shares when building a stock price prediction model for multiple companies. |
|
Keywords |
Stock price prediction, deep learning, LSTM, Tanzania |
|
URL |
