Project “Nontraditional regression models in transport modelling” (

Period: 01.10.2017
- 30.09.2020
Programme: ERAF


Assoc. professor 

Dr. sc. ing.
Nadezda Spiridovska
Engineering Faculty
Nadezda Spiridovska

Academic degree and current position in TSI: associate professor, researcher at the Engineering Faculty.

Previous experience: worked at the Mathematical Methods and Modelling Department in TSI for more than 10 years (assistant, lecturer, assistant professor), postdoc researcher.

Membership: A member of the Latvian Simulation Society since 2004, a member of the Latvian Operations Research Society (LatORS) since 2019.

Teaching activity: Discrete Mathematics (undergraduate level); Probability Theory and Mathematical Statistics (undergraduate level); System Modelling (undergraduate level); Introduction to machine learning (undergraduate level); Statistics (undergraduate level).

Publication activity: Author of on average two to three academic articles a year.

Projects: as a researcher has participated in more than 5 European and Latvian research projects.

Supervised Doctoral, Master and Bachelor Theses (number): 1 Master thesis and more than 10 Bachelor theses.

Research fields/domains: Data analysis, statistical analysis, mathematical modelling and simulation (in transportation), machine learning.

Motto: “Any problem can be turned into opportunity”.

Postdoc: Nadezda Spiridovska

The goal of the project is to develop nontraditional regression models, namely the Markov-modulated regression for analysis and forecasting of traffic flows and adjacent transport tasks in transport modelling, and find algorithms for their parameter estimation for big data.

The main objectives of the project are:

  • Estimation of the Markov-modulated linear regression parameters and forecasting of traffic flows on real data, taking into account the influence of the “external environment”
  • Development of the Markov-modulated linear regression model (multivariate regression, the case of a sample with missing data)
  • Development of algorithms for estimating Markov-modulated regression parameters on the basis of big data

Project publicity information:

contact us