• A
  • A
  • A
  • ABC
  • ABC
  • ABC
  • А
  • А
  • А
  • А
  • А
Regular version of the site

Educational Programme
Final Grade
Year of Graduation
Alexander Vorontsov
Financial Time Series: Multi-step Ahead Prediction
Data Science
(Master’s programme)
Master’s thesis on subject: “Time Series Forecasting: Multi-Step Ahead Prediction”.

Author: Vorontsov Alexander.

The given work consists of theoretical part, problem statement, algorithm implementation description and experiments. The experiments were made using artificial Lorenz time-series and real financial time series – BTC/USD rates from July to September of 2018.

In the theoretical part, there were given the main concepts and properties of chaotic time series and dynamical systems, generating it. Then, the main approaches to time-series forecasting were stated. After that, the thesis describes predictive clustering approach and some metrics to measure clustering quality. In this part of the thesis the author makes a decision to use graph-based Zahn clustering algorithm to cluster time series and perform forecasting.

In the second chapter the author brings mathematical problem statement of multi-step ahead prediction and predictive clustering. Following that, the metrics to measure forecasting quality were given.

The third part of given thesis is dedicated to Zahn algorithm implementation description and significances of applying it to Lorenz series and BTC/USD series. The algorithm was implemented in Python programming language. Them the author describes the hyperparameter optimization of implemented algorithm. This chapter also describes template forecasting technique and time series template realization.

In the fourth part the author gives detailed time series description and describes templates for each time series. In this part the thesis introduces forecasting method using prior information. For each time series the author describes experiments of using Zahn algorithm with or without prior information and measures errors. The author also brings a technique of using classifiers to predict Zahn algorithm’s error, that leads to forecast improvement.

In conclusion the thesis gives a summary of effectiveness of Zahn algorithm to perform multi-step ahead time-series forecast. The author gives recommendations of improving algorithm to make better predictions

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