ITDS | SPMS

MOBILE APP FOR STOCK MARKET PREDICTION USING LONG-SHORT-TERM MEMORY (LSTM)

Project Description
Supervisor

Market volatility and complexity are the inherent problems in the stock market, which further create problems for investors and traders regarding market unpredictability, incomplete information, and financial risk. Hence, the focus of the research is to develop a deep learning model in stock market prediction that can accurately forecast future trends in stocks. We have developed a robust predictive model by analyzing big amounts of historical stock data using some of the most advanced deep learning algorithms. Preprocessing of the data was conducted: imputation of missing values, normalization through Min-Max Scaling stock prices, and selection of the main feature, 'Close'; splitting the data into training, validation, and test sets. In this regard, the data analysis during exploration will focus on the following aspects: historical closing price visualization, analysis of the general volume of trading, and correlation analysis; assessment of daily returns in terms of pairwise relationships. Features such as daily return, VaR, and future price prediction were extracted from the data to enhance the predictive power of the model. We propose a comparative study between two models: LSTM and Linear Regression. The latter serves as our baseline. A variant of the LSTM model, very well known to learn long-term dependencies in time-series data, was implemented with an architecture that includes the stacking of LSTM units and dropout layers on top of each other and ending the model with a dense output layer. It tuned the hyperparameters on the LSTM model, changing the number of units in the hidden layer batch size, and epochs. The evaluation metrics showed a baseline model performing better than the proposed LSTM model. The baseline achieved an MSE value of 7.976, R² of 0.974, whereas, on the other side, the LSTM showed an MSE of 62.149 and R² of 0.798. Precision metrics were also in favor of the baseline model, standing at an impressive 98.67% as opposed to the LSTM model's 95.81%. In collaboration with Streamlit and Hugging Face, an amicable UI was developed that included interaction through buttons, menus, and icons for ease of use. This user interface is very important in terms of usability and the overall success of the digital product, whereby experts and non-experts alike can easily interact with the system. The project would not only lead to an improvement in the decision-making of individual investors but also prove useful for analysts and institutions. By advancing financial forecasting techniques, this research holds potential to significantly enhance the efficiency and accessibility of financial services globally, particularly in developing regions. The innovative approach of integrating a user-friendly app with a carefully trained LSTM model underscores the project's contribution to revolutionizing stock market prediction outcomes.

Machine Learning: Machine Learning (ML) research in Computer Science and Information Technology focuses on the development of algorithms and models that enable computers to learn from data and improve their performance over time without being explicitly programmed. It is a subset of Artificial Intelligence that uses statistical techniques to give machines the ability to learn patterns, make decisions, and predict outcomes based on data. Supervised learning, a key area of ML research, involves training models on labeled data, where the input-output relationships are predefined. This method is widely used for tasks such as classification (e.g., spam detection) and regression (e.g., predicting house prices). Unsupervised learning, on the other hand, involves finding hidden patterns in data without predefined labels, with clustering and association being typical applications in areas such as customer segmentation and anomaly detection. Reinforcement learning is another area of ML that focuses on teaching agents to make decisions by interacting with their environment and receiving feedback in the form of rewards or penalties. It is often applied in robotics, game playing, and autonomous systems, where continuous learning and adaptation are required.

The main objective is to design, develop, and implement a predictive model that utilizes LSTM to analyze historical stock market data and forecast trends.

2023/2024

Nov 17, 2024

HALIMATU SADIA MOHAMMED YUSSIF MARIAM EMMANUEL KWAKU BOAFO URIAS ESHUN BRACE BRIGHT APPIAGYEI MENSAH

BSc Information Technology