This project is a collection of predictive models and quantitative tools for stock price forecasting. It implements a variety of machine learning architectures, including generative adversarial networks, long short-term memory networks, and language models for financial analysis. The system distinguishes itself by combining time-series forecasting with natural language processing to convert financial news into numerical sentiment scores. It also incorporates synthetic market data generation and automated hyperparameter optimization using Bayesian and reinforcement learning methods to reduce p
Solution of the given task of predicting the buying and selling volume of the corporate bonds by treating it as a time series problem. The details of the solution and the techniques implemented can be found in Documentation.pdf, and bonds.ipynb and bonds_ts.ipynb respectively.