Joint Econometrics-Finance Lunch Seminar: Kenneth Kjær Jensen, AU

Title: Individualized housing indices using machine learning

2020.09.03 | Pernille Vorsø Jachobsen

Date Tue 17 Nov
Time 12:05 12:35
Location Online

Presenter: Kenneth Kjær Jensen, AU

Title: Individualized housing indices using machine learning

Abstract:”This paper proposes individualized housing indices for Danish housing units. The indices are constructed based on a population-wide housing dataset. In the methodology, principal components analysis (PCA) is tailored to include ordinal variables through the use of Pearson, polychoric, and polyserial correlation estimates. Furthermore, autoeconders, constituting a modern nonlinear counterpart of PCA, are employed to set contrast to the traditional PCA. Distinct sub-market divisions are obtained via autoencoders that match the true representation of the housing market. The index is then used in a pricing exercise. ”

The 'Accounting and Finance' and 'Econometrics and Business Statistics' Sections arrange lunch seminars on a regular basis. The speakers are usually section members, mostly PhD students and postdocs. The speakers can also be visiting PhD students or professors. Each lunch seminar is followed by a discussion of the paper presented.

Organizer: Daniel Borup

Joint Econometrics-Finance Lunch Seminars