News:
- April 2024: Here is the preprint Clustered Conformal Prediction for the Housing Market. This is submitted to the COPA 2024 conference, and is joint work with Johan Pensar and Jonathan P. Williams.
- December 2023: Here is the working paper Uncertainty quantification in automated valuation models with locally weighted conformal prediction. This is joint work with Gudmund Horn Hermansen, Johan Pensar, and Jonathan P. Williams.
- September 2023: Our paper, Locally Interpretable Tree Boosting: An Application to House Price Prediction, is accepted in Decision Support Systems' special issue on Explainable AI. You can read it here. Joint work with Ida Scheel, Dag Einar Sommervoll, and Johan Pensar.
- July 2023: Happy to announce that I will spend the fall semester as a Visiting Reseacher at the NCSU Statistics Department in Raleigh, North Carolina. I'll be working with Dr. Jonathan P. Williams and his group. Our research will be focused on conformal prediction applied to the housing market.
- January 2023: I just attended the Machine Learning Summer School (MLSS) outside of Cape Town, South Africa. Great experience!
- December 2022: Here is a quick tutorial on conformal prediction from BigInsight Day 2022.
- August 2022: I just presented a poster titled House Price Prediction with Confidence: Empirical Evidence from the Norwegian Market at the annual COPA symposium in Brighton (UK). The poster is here and the extended abstract is here.
- July 2022: Here is a work-in-progress titled Interpretable House Price Prediction Using a Collection of Local Machine Learning Models I recently presented at the WEAI 97th Annual Conference in Portland (Oregon) this summer. The corresponding slides are here.
- April 2022: I recently gave a short introduction to conformal inference to my supervisors. The introduction is my attempt at summarizing some parts of this book, this article and this article. My presentation can be found here.
- April 2022: Happy to announce that the paper House Price Prediction With Gradient Boosted Trees Under Different Loss Functions has been accepted to the Journal of Property Research. The paper is joint work with Johan Pensar, Ida Scheel, and Dag Einar Sommervoll. Full paper can be found here.
- July 2021: I just presented a working paper with the title House Price Prediction:
Classical Methods vs. Machine Learning Methods on the (virtual) WEAI 96th Annual Conference. The paper is joint work with Johan Pensar, Ida Scheel and Dag Einar Sommervoll. See the presentation here.
Academic interests:
- Tree-based machine learning methods
- Conformal prediction
- Statistical methods for house price prediction
- Computational statistics
Background:
I am an industrial PhD candidate at the Statistics and Data Science group at Department of Mathematics.
My PhD project (Uncertainty in house price prediction, 2021-2024) aims to study methods for house price prediction, with special interest in the uncertainty associated with a prediction. We are especially interested in uncertainty in tree based methods for supervised machine learning, such as random forest and boosted trees.
The project is supervised by associate professor Johan Pensar, associate professor Ida Scheel and professor Dag Einar Sommervoll (NMBU). The project is conducted in collaboration with Eiendomsverdi, a Norwegian property technology company.
I completed my MSc in Industrial Mathematics from NTNU in Trondheim in 2019, where I also spent one academic year at TU Berlin in Germany. You can find my master thesis here. I have previously worked as a consultant for Deloitte.