Universal Feature of Intraday Price Formation

This lecture by the instructor covers the application of deep learning techniques to high-frequency data, focusing on the universal feature of intraday price formation. The lecture delves into the use of neural networks for forecasting price changes based on order flow history, emphasizing the importance of universality in modeling stock behavior across different securities. The instructor presents results showing the effectiveness of pooling data from various stocks for forecasting accuracy, highlighting the role of machine learning in extracting relevant information from complex datasets. Additionally, the lecture explores the concept of stationarity in the relation between supply and demand fluctuations and price changes, demonstrating the stability and universality of this relation over time.