Experts’ Aggregation Algorithm for Long-Term Forecasting

Student: Korotin Alexander

Supervisor: Evgeny V. Burnaev

Educational Programme: Data Science (Master)

Year of Graduation: 2018

The article is devoted to investigating the application of aggregating algorithms to the problem of long-term forecasting, that is, prediction of the D-th step ahead outcome with experts’ advice. We develop the general aggregating algorithm G(D) based on the exponential reweighing of experts’ losses and prove its competitive loss bound. The designed algorithm can be applied to both finite and infinite(parametric) sets of experts. We show how the algorithm extends classical Vovk’s, Fixed Share and Mixing Past Posteriors methods to the setting of long-term forecasting. We obtain their long-term forecasting regret bounds, mainly sublinear of the game length T. In addition, we provide the application of the algorithm to the problem of online long-term linear regression.

Student Theses at HSE must be completed in accordance with the University Rules and regulations specified by each educational programme.

Summaries of all theses must be published and made freely available on the HSE website.

The full text of a thesis can be published in open access on the HSE website only if the authoring student (copyright holder) agrees, or, if the thesis was written by a team of students, if all the co-authors (copyright holders) agree. After a thesis is published on the HSE website, it obtains the status of an online publication.

Student theses are objects of copyright and their use is subject to limitations in accordance with the Russian Federation’s law on intellectual property.

In the event that a thesis is quoted or otherwise used, reference to the author’s name and the source of quotation is required.

Search all student theses