© Laurence Hervieux-Gosselin,

other photos: [ headshot ] [ old ]

Professeur

Université du Québec à Montréal

Machine learning, applied macroeconomics, time series econometrics and climate modeling.

Contact: p.gouletcoulombe [at] gmail [dot] com

[ Google Scholar ] [ CV ] [ Chaire Macro/Prévisions ] [ Linkedin ] [ Spotify ]

News

  • (23/11/19) New Paper The Anatomy of Machine Learning-Based Portfolio Performance is on SSRN. This is joint work with Dave Rapach, Erik Christian Montes Schütte, and Sander Schwenk-Nebbe.

  • (23/11/19) New Paper From Reactive to Proactive Volatility Modeling with Hemisphere Neural Networks is on SSRN. This is joint work with Mikael Frenette and Karin Klieber.

  • (23/11/19) L'émission « Format économique » a été diffusée sur la chaîne Savoir Média, où j'explique une série de concepts clés liés à l'inflation.

  • (23/05/22) Décorum’s song “Atterrissage” in now Top 3 on SiriusXM channel Racines Musicales. Next shows are June 16th at Festival Expé and June 22nd at Escogriffre in Montreal.

  • (23/05/11) I recently participated in the Radio-Canada “Question d’intérêt” podcast as a guest (in french)

  • (23/04/28) New Paper Maximally Machine-Learnable Portfolios is on SSRN. This is joint work with Maximilian Göbel.

  • (22/02/28) Some guitar solo from a recent show.

  • (22/11/29) New Album L’Abysse by my band Décorum is available on Spotify and all other streaming platforms.

  • (22/11/22) New Paper The Anatomy of Out-of-Sample Forecasting Accuracy is on SSRN. This is joint work with Daniel Borup, Dave Rapach, Erik Christian Montes Schütte, and Sander Schwenk-Nebbe. Check out the great blog post by Sander here. Python package here.

  • (22/06/21) New WebApp here we provide the daily UPenn-UQAM arctic sea ice forecasts of September 2022 based on a few models from recent papers (FELRs, FEMLs, and the VARCTIC). The user can investigate such predictions and those of previous years through a series of interactive plots.

  • (22/06/21) New Paper Assessing and Comparing Fixed-Target Forecasts of Arctic Sea Ice: Glide Charts for Feature-Engineered Linear Regression and Machine Learning Models is on Arxiv. This is joint work with Frank Diebold and Maximilian Göbel.

  • (22/06/16) First single from Décorum is out on Spotify and other platforms.

  • (22/06/03) Ryan Lucas and I just released a Python version of the Macro RF package. It has a few additional options like parallel computing and financial trading backtest functionalities. We made a webpage with a few toy applications — check it out.

  • (22/04/20) I’ll be visiting Bocconi University’s Department of Finance until the end of May.

  • (22/03/22) New paper When Will Arctic Sea Ice Disappear? Projections of Area, Extent, Thickness, and Volume is on Arxiv. This is joint work with Frank Diebold, Maximilian Göbel, Glenn Rudebusch, and Boyuan Zhang.

  • (22/02/10) My recent work on AI and inflation has been covered in the news. Some links (in French): Radio-Canada, TVA, Actualités UQAM, CScience.

  • (22/01/27) New paper A Neural Phillips Curve and a Deep Output Gap is on SSRN (and soon enough, on Arxiv).

  • (21/10/25) My new Album “Arrhythmia” is out on Spotify and other platforms. Don’t worry — I have not given up on econometrics just yet. Some interesting deep learning stuff is coming up soon.

  • (21/07/01) MRF, TVP Ridge, and Bag of Prunes packages can now be easily loaded directly in R.

  • (21/06/14) Check out this blog by Dave Wigglesworth. He uses Macro RF to forecast the Employment Cost Index over the last 5 years (including COVID) and finds that MRF provides an appreciably tight out-of-sample fit. The nowcast for 2020q2 (to be released July 30th) is 2.84, suggesting an upward trajectory for yet another price pressure indicator.

  • (21/06/09) A new version of To Bag is to Prune is on Arxiv. It now includes new experiments and a section on double descent and deep learning.

  • (21/06/01) A new… song is out!

  • (21/04/08) Successfully completed my thesis defense at Penn! [Slides]

  • (21/04/08) New R package TVPRidge implementing the ridge estimation approach for time-varying parameters (i.e., no more filtering, and cross-validation is already canned in). Check Code & Data for details and the original paper here.

  • (21/04/07) MRF package has been updated to version 0.1.1 which now includes a new function permitting to forecast given an already estimated MRF object. Among other things, this allows for iterated (multivariate) forecasts.

  • (21/03/23) New paper On Spurious Causality, CO2, and Global Temperature with Maximilian Göbel is on Arxiv.

  • (21/03/02) New paper Slow-Growing Trees is on Arxiv.

  • (21/03/02) Penn Grad Talks video is available here (mine starts at 26:13). It’s an almost accessible blitzkrieg overview of Macroeconomic Random Forests. [Slides]

  • (21/03/01) New paper Can Machine Learning Catch the COVID-19 Recession? with Massimiliano Marcelino and Dalibor Stevanovic is on Arxiv.

    • New large macro dataset for the UK available here.

  • (21/02/08) I’ll be joining UQÀM as Assistant Professor of Economics in the Fall!