SAIFIN

One of the most promising methods for evaluating trading strategies is based on the analysis of satellite data. The main goal of the project is to develop an algorithmic trading system based on Artificial Intelligence (AI) capable of identifying financial trading strategies by leveraging data and information available from the web and satellite sources. This approach, which utilizes satellite data and high-performance computing, enables nowcasting of key metrics to achieve better results in high-frequency trading (HFT), as well as in short-term and medium-to-long-term trading. The necessity of determining effective trading strategies also stems from the complexity of the dynamics driving financial markets, which makes it challenging to formulate successful strategies. Specifically, executing models based on large datasets with many variables requires computational power, parallelization, and efficiency. For this reason, the project will develop algorithms specifically designed for GPUs. The system will consist of two modules, which will form the two work packages of this project: the first module will handle the extraction of information from the web and satellite data, followed by processing to generate a synthetic and functional representation of the data. The second module will be the actual automated trading system, where the information gathered from the first module will be used to train an AI-based model aimed at determining an effective trading strategy.