The main aim of this project is to build a system that can predict stock prices using historical data and market trends by employing machine learning and deep learning techniques, specifically K-Nearest Neighbour (KNN) and Long Short-Term Memory (LSTM) networks. Accurate stock price forecasting is a challenging task due to the dynamic and non-linear nature of financial markets. This system seeks to assist investors and analysts by analyzing time-series data from various sources such as past stock performance, economic indicators, and domain-specific behavior to predict future trends. While perfect accuracy cannot be guaranteed, the project focuses on enhancing financial forecasting capabilities and helping students understand the intersection of artificial intelligence and finance.
The project will be executed over twelve weeks, with each stage contributing to the final predictive system. Initially, students will explore the fundamentals of machine learning models relevant to stock prediction. They will also research stock market dynamics, familiarize themselves with financial datasets, and gather historical stock price data along with relevant economic indicators.
Students will develop a basic framework using Python and libraries such as scikit-learn, TensorFlow, or Keras. The core implementation will involve building and training models using KNN for pattern recognition and LSTM for time series forecasting. After training, the models will be tested on real stock data to assess their prediction performance. Further tuning and optimization will be carried out to improve accuracy. In the final weeks, students will document their work, test the end-to-end system, and present the final outcome. Tools like Anaconda Navigator or Google Colab can be used for development, and stock visualization software may be included for better interpretation of predictions.
