High-frequency financial forecasting

Berkindale Analytics is a cloud-based company and wants to create a next-generation platform to provide both market data and packaged analytics to financial clients via direct download, interface, and dashboards accessible from a web portal. It is distinguished by its service arm that engages in data discovery, new analytics design, and ETL (“Extract, Transform, and Load”) configuration with client data.
The student will develop supervised and unsupervised machine learning tools adapted to capture stylized facts of a market or a trading platform, such as liquidity and volatility indicators. Berkindale is looking to develop a short-term (intraday) liquidity and volatility forecasting model for market indices and individual stocks. This model should use machine learning to combine forecasts obtained from different econometric models, depending on the situation (market conditions for related stocks, forecast events such as earnings announcements or macroeconomics, intensity of media coverage, etc.). This model will allow Berkindale’s clients to better manage their execution and volatility risk, thereby reducing their costs and improving their investment performance.

Faculty Supervisor:

Ioannis Mitliagkas

Student:

Partner:

Berkindale Analytics Inc

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

Université de Montréal

Program: