Personalized Wealth Management Advisor based on the Analysis of Times Series Related to Financial Transactions

The aim of this research project is to develop innovative tools that will help financial institutions deliver highly personalized services to their customers. We intend to use the most recent advances in statistical learning methods and machine learning algorithms mostly in deep learning, vector embeddings and autoencoders, to leverage the power of time series models by extracting high-level features from both assets and customers’ transactional data. In order to build the required innovative tools, we will adapt existing machine learning algorithms and develop new ones that will be fed with the extracted high-level features to assess customers situations and match them with financial strategies. Comfiz has been working in this domain with leading Canadian financial institutions for the past 18 months and as identified several areas where leveraging time series analysis could lead to significant improvements in the capacity of financial institutions to provide more personalized and dynamic recommendations to their clients.

Intern: 
Benjamin Choteau
Quentin Golliot
Faculty Supervisor: 
Luc Adjengue
Province: 
Quebec
Program: